Data processing method and system

Abstract

The data processing method and system according to the present specification can divide an initial frame into multiple units and adjust the amplitude in the mid- to high-frequency region of each unit using different edge adjustment coefficients. If the amplitude in the mid- to high-frequency region of the current unit is large, an edge adjustment coefficient smaller than 1 and larger than 0 is used to reduce the amplitude in the mid- to high-frequency region, thereby reducing the amount of data information and improving the efficiency of data compression. If the amplitude in the mid- to high-frequency region of the current unit is small, an edge adjustment coefficient larger than 1 is used to enhance the amplitude in the mid- to high-frequency region to avoid detail loss. In the case of data decompression, the method and system can perform edge compensation on the amplitude in the mid- to high-frequency region of each unit using an edge compensation coefficient corresponding to the edge adjustment coefficient, and restore the edge-adjusted compressed data to the clarity of the initial frame or a clarity higher than that of the initial frame, thereby obtaining a decompressed frame.

Description

[Technical Field]

[0001] The present specification relates to the field of data processing, and in particular to a data processing method and system. [Background technology]

[0002] With the spread of Internet technology, especially the spread of mobile terminals, more and more types of data are appearing in communication networks, and with the spread of computers, more and more data, such as video data, audio data, etc., are occupying more and more network and storage resources. Since data often contains a huge amount of information and has high storage and transmission requirements, data is often compressed to facilitate storage and transmission, and the compressed data is often decompressed and restored when needed. Therefore, data compression and decompression technologies are being applied more and more.

[0003] For example, video and image compression technologies have been increasingly applied in recent decades. Video often contains a huge amount of information. Compressed video and images, from traditional radio, movie, and television broadcasting to current mass surveillance and Internet applications, increasingly occupy network and storage resources. As a result, transmitting raw video data from one terminal to another over a network occupies a large amount of network resources. This can make smooth transmission of images difficult in real-time video transmission. Therefore, to facilitate transmission, video data must first be compressed by a data compression device before transmission. After the compressed video is transmitted over a transmission medium to a data decompression device, the data decompression device decompresses the video to at least partially restore the video image.

[0004] The main video compression standards in the prior art are the H.264 and H.265 standards. Before transmission, video is typically globally compressed in an encoder according to the H.264 and H.265 standards, and then globally decompressed in a decoder according to the H.264 and H.265 standards after transmission. However, the above-described global video compression method is unsatisfactory in terms of the balance between computational complexity and video clarity after decompression. This is because the H.264 and H.265 standards require various complex algorithms to generate a predicted frame of the original frame when processing the original video, and then record the residual between the original frame and the predicted frame. The closer the predicted frame is to the original frame, the smaller the residual, resulting in a smaller amount of data after video encoding. To facilitate encoding, a common method is to reduce high-frequency information in the original frame image by filtering the original frame. As can be seen from Fourier transforms, the edge of an object in an image often has rich frequency information, and the high-frequency components of the edge are usually larger than those of other gentle areas. As a result, frame images with reduced high-frequency information appear visually blurred (i.e., image clarity is reduced), but the residual error between the predicted frame and the filtered original frame can be reduced. This significantly reduces both the amount of computation required for video encoding and the encoded data stream. However, frame prediction technology is very complex and consumes a large amount of computational resources. Taking a video encoding / decoding system as an example, on average, every 30% to 40% improvement in coding efficiency increases the amount of computation by approximately 10 times. Furthermore, transmitted data often loses clarity after being decompressed, and various noise artifacts, such as blocking artifacts or ringing artifacts, often occur. Blocking artifacts refer to discontinuities at image edges caused by block-based Fourier transforms in image processing.The ringing artifact refers to the phenomenon that, when performing a spectral adjustment process on an image, if there is a large change in the value of the selected spectral adjustment function (i.e., there is a region where the derivative changes rapidly), gray level oscillations occur in the output image in areas where gray levels change rapidly, similar to the air vibrations that occur after a bell is struck. This noise is particularly prevalent at image edges. If an output image contains strong noise, it cannot meet people's ever-increasing demand for data clarity. Therefore, further improving data compression efficiency, improving clarity after data decompression, and eliminating noise are sought after in the data compression and decompression technology fields.

[0005] Therefore, in order to improve the efficiency of data transmission and the clarity of data after decompression, there is a need for a data processing method and system that has higher compression efficiency and clearer data decompression. Summary of the Invention

[0006] This specification provides a data processing method and system with higher compression efficiency and clearer data decompression. Taking video data as an example, the data processing method and system can divide an initial frame of initial video data into multiple units, obtain the amplitude of the mid- to high-frequency region of each unit, and adjust the amplitude of the mid- to high-frequency region of each unit using different edge adjustment coefficients to reduce the amplitude of the mid- to high-frequency region of the initial frame. If the amplitude of the mid- to high-frequency region of a current unit is large, indicating that the current unit contains a strong edge, the edge adjustment coefficient is less than 1 and greater than 0 to adjust the amplitude of the mid- to high-frequency region of the current unit, thereby reducing the signal strength in the mid- to high-frequency region of the current unit, reducing the amount of data information, and improving the efficiency of data compression when predicting and obtaining a residual. If the amplitude of the mid-frequency to high-frequency region of the current unit is small, indicating that the current unit contains a weak edge, an edge adjustment coefficient greater than 1 is used to adjust the amplitude of the mid-frequency to high-frequency region of the current unit, thereby increasing the amplitude of the mid-frequency to high-frequency region of the current unit, thereby preventing the weak edge in the current unit from being lost during data compression (prediction and determination of residual) and avoiding detail loss.The data processing method and system improve the efficiency of data compression and increase the data information amount of weak edges, thereby preventing detail loss during the data compression process, i.e., improving data compression efficiency and reducing data distortion.

[0007] When performing data decompression on a compressed frame, the method and system use the data compression units as data decompression units and perform edge compensation on the amplitude of the mid- to high-frequency range of each unit using edge compensation coefficients corresponding to edge adjustment coefficients to compensate for the amplitude loss in the mid- to high-frequency range during the data compression process, thereby obtaining a decompressed frame. The edge compensation corresponds to the edge adjustment, and there is a corresponding relationship between the edge compensation coefficients and the edge adjustment coefficients. The edge compensation can restore the edge-adjusted compressed data to the clarity of the initial frame or a clarity higher than that of the initial frame. In other words, the decoding side can restore data within at least the important frequencies of the decompressed data to the clarity of the initial frame without significantly increasing the amount of encoding and decoding calculations, and can even obtain a clarity higher than that of the initial frame.

[0008] Since the edge adjustment coefficients in the edge adjustment process of the initial frame are all greater than 0 and no information is lost in the compressed frame, the edge adjustment coefficients and edge compensation coefficients can be designed based on the relationship between the edge adjustment coefficients and the edge compensation coefficients and their respective characteristics to restore the information in the compressed frame.The method and system can significantly improve data compression efficiency, improve data transmission efficiency, reduce data loss, avoid detail loss, remove noise, and improve the clarity of the decompressed data.

[0009] Based on this, according to a first aspect, the present specification provides a data processing method, the method including: selecting an initial frame in initial data, the initial frame including a predetermined number of bytes of initial data; and performing data compression on the initial frame to obtain a compressed frame, wherein the data compression includes performing edge adjustment on the frame being compressed, the frame being compressed including the initial frame and any data state of the initial frame before it becomes the compressed frame in the data compression process, and wherein the edge adjustment includes adjusting the amplitude in the mid-frequency to high-frequency region of each unit among a plurality of units of the frame being compressed using a corresponding edge adjustment coefficient, thereby reducing the amplitude in the mid-frequency to high-frequency region of the frame being compressed, and the edge adjustment coefficient is greater than 0.

[0010] In some embodiments, the edge adjustment performed on the compressed frame as described above includes dividing the compressed frame into the plurality of units based on a predetermined unit size, and adjusting the amplitude of each unit in the mid- to high-frequency range using the corresponding edge adjustment coefficients.

[0011] In some embodiments, adjusting the amplitude of each unit in the mid-frequency to high-frequency range using the edge adjustment coefficient corresponding thereto as described above includes: selecting one function from a group of preset encoding functions as an encoding function for each unit; adjusting each unit using the encoding function to obtain a first unit so that components in the low-frequency range within the frequency range are maintained while components in the mid-frequency to high-frequency range are attenuated; calculating the difference between each unit and the first unit to obtain a first edge corresponding thereto, wherein the first edge includes the components in the mid-frequency to high-frequency range; adjusting the amplitude of the first edge using the edge adjustment coefficient corresponding thereto to obtain a corresponding encoded edge; and superimposing the first unit and the encoded edge.

[0012] In some embodiments, adjusting the amplitude of the first edge using the corresponding edge adjustment factor as described above includes determining that the edge value of the first edge is less than a predetermined first threshold and increasing the amplitude of the first edge by the edge adjustment factor greater than 1, or determining that the edge value of the first edge is greater than a predetermined second threshold and decreasing the amplitude of the first edge by the edge adjustment factor less than 1.

[0013] In some embodiments, the aforementioned step of increasing the amplitude of the first edge by the edge adjustment factor greater than 1 includes selecting a coefficient from a preset first set of edge adjustment coefficients as the edge adjustment factor to increase the amplitude of the first edge, where all coefficients in the first set of edge adjustment coefficients are greater than 1; and the aforementioned step of reducing the amplitude of the first edge by the edge adjustment factor less than 1 includes selecting a coefficient from a preset second set of edge adjustment coefficients as the edge adjustment factor to reduce the amplitude of the first edge, where all coefficients in the second set of edge adjustment coefficients are less than 1.

[0014] In some embodiments, adjusting the amplitude of the first edge using the corresponding edge adjustment coefficients as described above includes: obtaining the edge adjustment coefficients based on an optimization algorithm using weighted values ​​of a distortion rate and a coding rate as optimization targets; and adjusting the amplitude of the first edge using the edge adjustment coefficients.

[0015] In some embodiments, the aforementioned data compression on the initial frame includes at least one of a method of first performing the edge adjustment on the initial frame, and then predicting the edge-adjusted initial frame to obtain a residual; a method of first performing prediction on the initial frame to obtain a predicted frame, and then performing the edge adjustment on the initial frame and the predicted frame to obtain a residual; and a method of first performing prediction on the initial frame, obtaining a residual, and then performing the edge adjustment on the residual.

[0016] In some embodiments, the compressed frame further comprises the encoding function and the edge adjustment coefficient corresponding to each unit of the plurality of units.

[0017] According to a second aspect, the present specification further provides a data processing system, the system including at least one storage medium and at least one processor, the at least one storage medium storing at least one instruction set for processing data, the at least one processor being communicatively connected to the at least one storage medium, wherein when the system is operating, the at least one processor reads the at least one instruction set and performs the data processing method described in the first aspect of the present specification in accordance with instructions in the at least one instruction set.

[0018] According to a third aspect, the present specification further provides a data processing method, the method including: obtaining compressed data, the compressed data including a compressed frame obtained by performing data compression on an initial frame, the data compression including edge adjustment; and performing data decompression on the compressed frame to obtain a decompressed frame, the data decompression including performing edge compensation on a frame being decompressed, the frame being decompressed including the compressed frame and any data state of the compressed frame before it becomes the decompressed frame in the data decompression process, wherein the edge compensation and the edge adjustment have a predetermined correlation.

[0019] In some embodiments, the edge adjustment includes adjusting the amplitude of each unit of the plurality of units of the frame being compressed in the mid-frequency to high-frequency region using a corresponding edge adjustment coefficient to reduce the amplitude of the frame being compressed in the mid-frequency to high-frequency region, the edge adjustment coefficient being greater than 0, the frame being compressed including the initial frame and any data state of the initial frame before it became the compressed frame in the data compression process, and the edge compensation includes compensating the amplitude of each unit of the plurality of units of the frame being decompressed in the mid-frequency to high-frequency region using an edge compensation coefficient corresponding to the edge adjustment coefficient based on the correlation.

[0020] In some embodiments, performing edge adjustment on the frame being compressed as described above includes dividing the frame being compressed into the plurality of units based on a predetermined unit size, and adjusting the amplitude of each unit in a mid- to high-frequency region using the edge adjustment coefficients corresponding to the units, wherein for each unit, one function is selected from a predetermined group of encoding functions as an encoding function and the encoding function is used to obtain a first unit so that components in the low-frequency region within the frequency region are maintained while components in the mid- to high-frequency region are attenuated; calculating a difference between the first unit and the first unit to obtain a first edge corresponding to the first unit, wherein the first edge includes the components in the mid- to high-frequency region; adjusting the amplitude of the first edge using the edge adjustment coefficients corresponding to the first unit to obtain a corresponding coded edge; and superimposing the first unit and the coded edge.

[0021] In some embodiments, performing edge compensation on the decompressed frame as described above includes dividing the decompressed frame into the plurality of units based on the predetermined unit size, and compensating the amplitude of each unit in the mid- to high-frequency range using the edge compensation coefficient corresponding to the edge adjustment coefficient.

[0022] In some embodiments, compensating the amplitude of each unit in the mid- to high-frequency range using the edge compensation coefficient corresponding to the edge adjustment coefficient as described above includes: determining a decoding function for each unit, adjusting it according to the decoding function to obtain a second unit such that components in the low-frequency range within the frequency range are retained while components in the mid- to high-frequency range are attenuated; calculating the difference between the second unit and the second unit to obtain a corresponding second edge, wherein the second edge includes the components in the mid- to high-frequency range; compensating the amplitude of the second edge using the edge compensation coefficient corresponding to the edge adjustment coefficient to obtain a corresponding decoded edge; and superimposing the decoded edge with the current unit.

[0023] In some embodiments, determining the decoding function as described above includes selecting a function from a group of predefined decoding functions as the decoding function.

[0024] In some embodiments, compensating the amplitude of the second edge using the edge compensation coefficient corresponding to the edge adjustment coefficient as described above includes selecting one coefficient from a group of preset edge compensation coefficients as the edge compensation coefficient to compensate the amplitude of the second edge.

[0025] In some embodiments, the compressed frame includes the encoding function and the edge adjustment coefficient corresponding to each unit of a plurality of units in the frame being compressed.

[0026] In some embodiments, determining the decoding function as described above includes selecting a function corresponding to the encoding function from a preset group of decoding functions as the decoding function.

[0027] In some embodiments, compensating the amplitude of the second edge using the edge compensation coefficient corresponding to the edge adjustment coefficient as described above includes determining the edge compensation coefficient based on a correlation between the edge adjustment coefficient and the edge compensation coefficient to compensate the amplitude of the second edge.

[0028] In some embodiments, performing data decompression on the compressed frame as described above includes at least one of the following methods: first decoding the compressed frame and then performing the edge compensation; performing the edge compensation while decoding the compressed frame; and first performing the edge compensation and then decoding the compressed frame.

[0029] In some embodiments, the correlation includes ensuring that the edge compensation results in an amplitude of the extended frame at any frequency in a low to mid frequency range that is 85% or greater than that of the initial frame.

[0030] In some embodiments, the correlation further comprises: the edge compensation causing a gentle increase in amplitude in the mid-frequency region of the extended frame relative to the initial frame.

[0031] In some embodiments, the correlation further comprises the edge compensation causing a gentle reduction in amplitude in high frequency regions of the extended frame relative to the initial frame.

[0032] According to a fourth aspect, the present specification further provides a data processing system, the system including at least one storage medium and at least one processor, the at least one storage medium storing at least one instruction set for processing data, the at least one processor being communicatively connected to the at least one storage medium, wherein when the system is operating, the at least one processor reads the at least one instruction set and performs the data processing method described in the third aspect of the present specification in accordance with instructions in the at least one instruction set.

[0033] Other features of the data processing method and system according to the present specification are partially described in the following description. According to the description, the contents shown in the following figures and examples will be obvious to those skilled in the art. The creative aspects of the data processing method, system and storage medium according to the present specification can be fully understood by practicing or using the methods, apparatuses and combinations described in the following detailed examples. [Brief explanation of the drawings]

[0034] In order to more clearly explain the technical solutions in the embodiments of this specification, the following will briefly describe the drawings that need to be used in the description of the embodiments. It should be obvious that the drawings in the following description are only some embodiments of this specification, and those skilled in the art can obtain other drawings based on these drawings without any creative efforts. [Figure 1] 1 is a schematic diagram of a data processing system according to an embodiment of the present disclosure; [Figure 2] 1 is a schematic diagram of a data compression device for data processing according to an embodiment of the present disclosure; [Figure 3A] 1 is a flowchart of data compression and data decompression according to an embodiment of the present disclosure. [Figure 3B] 1 is a flowchart of data compression and data decompression according to an embodiment of the present disclosure. [Figure 3C] 1 is a flowchart of data compression and data decompression according to an embodiment of the present disclosure. [Figure 4A] 1 is a flowchart of a data processing method for compressing data according to an embodiment of the present disclosure. [Figure 4B] 1 is a flowchart of edge adjustment according to an embodiment of the present disclosure. [Figure 5] FIG. 1 is a structural block diagram of edge adjustment according to an embodiment of the present disclosure. [Figure 6] 10 is a graph of an encoding function according to an embodiment of the present disclosure. [Figure 7A] 1 is a flowchart of a data processing method for decompressing a compressed frame, according to an embodiment of the present disclosure. [Figure 7B] 1 is a flowchart of edge compensation according to an embodiment of the present disclosure. [Figure 8] 1 is a structural flowchart of edge compensation according to an embodiment of the present disclosure; [Figure 9A] 10 is a graph of a global conditioning function H0(f) according to an embodiment of the present disclosure. [Figure 9B] 10 is a graph of a global conditioning function H0(f) according to an embodiment of the present disclosure. [Figure 9C] 10 is a graph of a global conditioning function H0(f) according to an embodiment of the present disclosure. [Figure 9D] 10 is a graph of a global conditioning function H0(f) according to an embodiment of the present disclosure. [Figure 10A] 10 is a graph of the overall adjustment function H0(f), the edge adjustment function H1(f), and the decoding function H2(f) of the normal mode according to an embodiment of the present disclosure. [Figure 10B] 10 is a graph of the overall adjustment function H0(f), the edge adjustment function H1(f), and the decoding function H2(f) of the enhancement mode according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0035] The following description provides specific application scenarios and requirements of the present specification to enable those skilled in the art to make and use the content herein. Various modifications to the disclosed embodiments will be apparent to those skilled in the art, and the general principles defined herein may be applied to other embodiments and applications without departing from the spirit and scope of the present specification. Therefore, the present specification is not intended to be limited to the embodiments shown, but is accorded the widest scope consistent with the claims.

[0036] The terminology used herein is for the purpose of describing particular example embodiments only and is not intended to be limiting. For example, the singular forms "a," "an," and "the" as used herein can also include the plural unless the context clearly dictates otherwise. As used herein, the terms "comprise," "include," and / or "containing" refer to the presence of associated integers, steps, operations, elements, and / or components, but do not exclude the presence of one or more other features, integers, steps, operations, elements, components, and / or groups, or that other features, integers, steps, operations, elements, components, and / or groups may be added to the system / method.

[0037] These and other features of the present specification, as well as the operation and function of the associated elements of structure, and the economies of assembly and manufacture of parts, can be clearly improved upon consideration of the following description. Reference is made to the drawings, all of which form a part of this specification. It is to be expressly understood, however, that the drawings are for the purpose of illustration and description only and are not intended to limit the scope of this specification. It should also be understood that the drawings are not drawn to scale.

[0038] As used herein, flowcharts illustrate operations of a system implementation according to some embodiments of the present invention. It should be clearly understood that the operations of a flowchart may be implemented out of order. Conversely, operations may be implemented in reverse order or simultaneously. It should be noted that one or more other operations may be added to a flowchart. One or more operations may be removed from a flowchart.

[0039] In one aspect, this specification provides a data processing system 100 (hereinafter abbreviated as system 100). In a second aspect, this specification describes a data processing method P200 for compressing data, and in a third aspect, this specification describes a data processing method P300 for decompressing compressed frames.

[0040] 1 shows a schematic diagram of a data processing system 100. The system 100 may include a data compression device 200, a data decompression device 300, and a transmission medium 120.

[0041] The data compression device 200 is capable of receiving an initial frame of initial data to be compressed and compressing the initial data to generate a compressed frame using the data processing method P200 proposed herein. The data compression device 200 is capable of storing and executing data and / or instructions for carrying out the data processing method P200 described herein.

[0042] The data decompression device 300 receives compressed frames and decompresses the compressed frames using the data processing method P300 proposed herein to obtain decompressed frames. The data decompression device 300 is capable of storing and executing data and / or instructions for carrying out the data processing method P300 described herein.

[0043] The data compression device 200 and the data decompression device 300 may include a wide range of devices, such as desktop computers, mobile computing devices, notebook (e.g., laptop) computers, tablet computers, set-top boxes, handsets such as smartphones, televisions, cameras, display devices, digital media players, video game consoles, in-vehicle computers, and the like.

[0044] As shown in FIG. 1 , the data compression device 200 and the data decompression device 300 may be connected via a transmission medium 120. The transmission medium 120 may facilitate the transmission of information and / or data. The transmission medium 120 may be any data carrier capable of transmitting compressed frames from the data compression device 200 to the data decompression device 300. For example, the transmission medium 120 may be a storage medium (e.g., an optical disk), a wired or wireless communication medium, or a network. In some embodiments, the transmission medium 120 may be any type of wired or wireless network, or a combination thereof. For example, the transmission medium 120 may include a cable network, a wired network, an optical fiber network, a telecommunications network, an intranet, the Internet, a local area network (LAN), a wide area network (WAN), a wireless local area network (WLAN), a metropolitan area network (MAN), a wide area network (WAN), a public switched telephone network (PSTN), a Bluetooth network, a ZigBee network, a near field communication (NFC) network, etc. One or more components in the data decompression device 300 and the data compression device 200 may be connected to a transmission medium 120 to transmit data and / or information. The transmission medium 120 may include a router, a switch, a base station, or other device that facilitates communication from the data compression device 200 to the data decompression device 300. In another embodiment, the transmission medium 120 may be a storage medium, such as mass memory, removable memory, volatile read-write memory, read-only memory (ROM), etc., or any combination thereof. Exemplary mass memory may include non-transitory storage media such as magnetic disks, optical disks, solid-state drives, etc. Removable memory may include flash drives, floppy disks, optical disks, memory cards, zip magnetic disks, magnetic tape, etc. Exemplary volatile read-write memory may include random access memory (RAM).The RAM may include dynamic RAM (DRAM), double data rate synchronous dynamic RAM (DDR SDRAM), static RAM (SRAM), thyristor RAM (T-RAM), zero capacitor RAM (Z-RAM), etc. The ROM may include masked read only memory (MROM), programmable read only memory (PROM), erasable programmable read only memory (PEROM), electrically erasable programmable read only memory (EEPROM), optical disk (CD-ROM), digital versatile disk ROM, etc. In some embodiments, the transmission medium 120 may be a cloud platform. By way of example only, the cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an intercloud, or similar forms, or any combination of these forms.

[0045] As shown in FIG. 1, the data compression device 200 receives initial data and executes instructions of a data processing method P200 described in this specification to perform data compression on the initial data to generate compressed frames, which are transmitted to the data decompression device 300 via the transmission medium 120, and the data decompression device 300 executes instructions of a data processing method P300 described in this specification to perform data decompression on the compressed frames to obtain decompressed frames.

[0046] 2 shows a schematic diagram of a data processing apparatus 200 for compressing data. The data processing apparatus 200 is capable of executing a data processing method P200 described herein, which is introduced elsewhere in this specification.

[0047] 2, the data compression device 200 includes at least one storage medium 230 and at least one compression-side processor 220. In some embodiments, the data compression device 200 may further include a communication port 250 and an internal communication bus 210. Additionally, the data compression device 200 may further include an I / O component 260.

[0048] The internal communication bus 210 may be connected to various system components, including a storage medium 230 and a compression processor 220 .

[0049] The I / O component 260 supports input / output between the data compression device 200 and other components.

[0050] Storage medium 230 may include a data storage device. The data storage device may be a non-transitory storage medium or a transitory storage medium. For example, the data storage device may include one or more of a magnetic disk 232, a read-only storage medium (ROM) 234, or a random access storage medium (RAM) 236. Storage medium 230 further includes at least one set of instructions stored on the data storage device. The instructions may be computer program code, which may include programs, routines, objects, components, data structures, processes, modules, etc., that perform data processing methods according to the present disclosure.

[0051] The communication port 250 is used for data communication between the data compression device 200 and the outside. For example, the data compression device 200 can be connected to the transmission medium 120 via the communication port 250.

[0052] The at least one compression processor 220 is communicatively connected to the at least one storage medium 230 via the internal communication bus 210. The at least one compression processor 220 is for executing the at least one instruction set. When the system 100 operates, the at least one compression processor 220 reads the at least one instruction set and executes the data processing method P200 according to the instructions of the at least one instruction set. The compression processor 220 can execute all steps included in the data processing method P200. The compression-side processor 220 may be in the form of one or more processors, and in some embodiments, the compression-side processor 220 may include one or more hardware processors, such as a microcontroller, a microprocessor, a reduced instruction set computer (RISC), a special-purpose integrated circuit (ASIC), an application-specific instruction set processor (ASIP), a central processing unit (CPU), a graphics processing unit (GPU), a physical processing unit (PPU), a microcontroller unit, a digital signal processor (DSP), a field-programmable gate array (FPGA), an advanced RISC machine (ARM), a programmable logic device (PLD), any circuit or processor capable of performing one or more functions, or any combination thereof. For illustrative purposes only, only one compression-side processor 220 is described herein for the data compression apparatus 200. However, it should be noted that the data compression apparatus 200 herein may further include multiple processors, and thus, the operations and / or method steps disclosed herein may be performed by a single processor or jointly by multiple processors as described herein.For example, in this specification, when the compression side processor 220 of the data compression device 200 performs step A and step B, it should be understood that step A and step B may be performed jointly or separately by two different compression side processors 220 (e.g., a first processor performs step A and a second processor performs step B, or the first and second processors perform steps A and B jointly).

[0053] Although the above structure is described for a data compression device 200, this structure is also applicable to a data decompression device 300. The data decompression device 300 is capable of executing a data processing method P300 described herein, which is introduced elsewhere in this specification.

[0054] The data processing methods P200, P300, and system 100 are used to compress and decompress data, thereby improving data transmission efficiency and saving resources and space. The data may be non-real-time data or real-time data. Various types of data exist, from traditional radio, movie, and television broadcasts to current mass surveillance and Internet applications. For example, the data may be non-real-time video data, audio data, or image data. The data may also be real-time map data, real-time sensor data, real-time video surveillance data, network observation data, weather data, aerospace data, and the like. For example, the data may be map data received by an autonomous vehicle from a base station while traveling. This specification does not limit the specific category of data. The data processing methods P200, P300, and system 100 described herein share the same methods and steps for processing different categories of data. For ease of explanation, this specification will describe processing of video data as an example.

[0055] Data compression and decompression are often performed frame-by-frame. A frame is a processing unit that constitutes a data sequence. The initial data may include one or more initial frames. Each initial frame includes a predetermined number of bytes of initial data. In video compression, the initial data may be initial video data, and the initial frame may be a frame image in the initial video data. Conventional video compression technologies typically achieve the goal of compressing the video data by encoding the initial video data using the H.264 and H.265 standards. When encoding video data using the H.264 and H.265 standards, predictive coding is mainly used, i.e., predicting the initial data in the video data to obtain a predicted value, and then subtracting the predicted value from the initial value of the initial data to obtain a residual value, thereby compressing the video data. During decompression and decompression (i.e., decoding), the initial frame can be restored by adding the residual value to the predicted value.

[0056] The data processing methods P200, P300 and system 100 according to this specification combine edge adjustment and encoding when performing data compression, thereby reducing the amount of data required for encoding and improving the efficiency of video data compression and video transmission. When performing data decompression, they combine edge compensation and decoding to decompress compressed data that has been edge adjusted and encoded, thereby restoring the decompressed data to its original data.

[0057] The data processing method P200 can perform data compression on the video data. The data processing method P200 can perform data compression on an initial frame using a method that combines encoding (i.e., prediction and residual calculation) and edge adjustment to obtain a compressed frame. Specifically, the data processing method P200 can perform the edge adjustment and the encoding on a currently compressed frame. The currently compressed frame includes the initial frame and any data state of the initial frame before it becomes the compressed frame in the data compression process. The edge adjustment refers to adjusting the amplitude of a spectrogram of the data to be processed. For example, the edge adjustment can adjust the amplitude of a selected region within the frequency domain of the data to be processed, such as the amplitude of a mid-frequency region, the amplitude of a high-frequency region, or further, for example, the amplitude of a low-frequency to mid-frequency region, or further, for example, the amplitude of a mid-frequency to high-frequency region. In some embodiments, the edge adjustment can reduce the amount of data information in the data to be processed by attenuating the amplitude of a selected region within the frequency domain using an edge adjustment coefficient greater than 0 and less than 1. As will be understood by those skilled in the art, the amplitude-attenuated data has smaller frequency components within the selected frequency range, and the amount of information in the data is reduced, which can improve the efficiency of encoding the amplitude-attenuated data and improve the compression rate.

[0058] The data processing method P300 can restore the video data by performing data decompression on the compressed frame that has been subjected to the data compression by the data processing method P200 to obtain a decompressed frame. The data processing method P300 can restore the data in the compressed frame by performing data decompression on the compressed frame using a method that combines decoding (i.e., restoring the frame being compressed based on a residual value and a predicted value) and edge compensation. Specifically, the data processing method P300 can perform the edge compensation and the decoding on the frame being decompressed. The frame being decompressed may include the compressed frame and any data state of the compressed frame before it became the decompressed frame in the data decompression process. The edge compensation can completely or approximately restore the edge-adjusted data to its state before edge adjustment, without considering other calculation errors.

[0059] Therefore, the data processing methods P200, P300 and system 100 can significantly improve the compression efficiency of video data, reduce data loss during video data compression, improve the video transmission efficiency, restoration rate, and clarity of the decompressed video, and reduce noise in the decompressed video. The specific processes of the edge adjustment and edge compensation will be described in detail later.

[0060] When the system 100 performs data compression on video data, the order of the edge adjustment and the encoding may be interchanged or crossed. The edge adjustment may be performed before or after the encoding. Similarly, when the system 100 performs data decompression on a compressed frame, the order of the edge compensation and the decoding may be interchanged or crossed. It should be noted that, to ensure that the decompressed data information can restore the information in the initial data, the order of the edge compensation and the decoding in the data decompression should correspond to the order of the edge adjustment and the encoding in the data compression; that is, the edge compensation and the decoding can be performed symmetrically in the reverse direction to the edge adjustment and the encoding. For example, if the compressed frame is obtained by performing the edge adjustment and then the encoding, when the compressed frame is decompressed, the decoding should be performed first, and then the edge compensation should be performed. For ease of explanation, the inventors define the data in the initial frame before data compression processing as P0, and the data in the decompressed frame obtained by decompression by the data decompression device 300 as P4.

[0061] In the data processing method P200, when performing data compression on the initial frame, the data compression device 200 may perform the edge adjustment on the initial frame before performing the encoding, or may perform the encoding on the initial frame before performing the edge adjustment. Figures 3A to 3C show flowcharts of several data compression and data decompression processes according to embodiments of the present specification.

[0062] FIG. 3A illustrates a flowchart of data compression and data decompression according to an embodiment of the present disclosure. As shown in FIG. 3A, the data compression device 200 may compress data on initial data as follows: The data compression device 200 first performs the edge adjustment on the initial frame P0, then performs the encoding, i.e., performs prediction on the edge-adjusted initial frame to obtain a residual, obtains predicted data PI and residual data R, and inputs the predicted data PI and the residual data R to a bitstream generation module for synthesis to obtain the compressed frame. The compressed frame includes the predicted data PI and the residual data R. Here, the frame being compressed may be the initial frame P0. For ease of explanation, the inventors define the edge-adjusted data as an encoding-adjusted frame P1. As described above, the edge adjustment may involve dividing the frame being compressed (initial frame P0) into multiple units and performing the edge adjustment on each unit using a corresponding edge adjustment coefficient. In some embodiments, the data compression device 200 may further input an encoding function and an edge adjustment coefficient corresponding to each unit in the edge adjustment to the bitstream generation module for synthesis. That is, the compressed frame may further include the encoding function and the edge adjustment coefficient corresponding to each unit in the edge adjustment. For ease of explanation, the inventors define the data of the encoding function and the edge adjustment coefficient corresponding to each unit in the compressed frame as encoded data RAMI (Regional Amplitude Modulation Information). For ease of explanation, in FIG. 3A, the inventors will use an example in which the compressed frame includes encoded data RAMI. The encoding function and the edge adjustment coefficient in the edge adjustment will be described in detail later. The data compression method shown in FIG. 3A can improve encoding efficiency, further reduce the amount of data in the compressed frame, improve the compression rate, reduce data loss, and avoid detail loss.

[0063] As shown in FIG. 3A, the data decompression device 300 may perform data decompression on the compressed frame as follows: the data decompression device 300 performs the decoding on the compressed frame and then performs the edge compensation. Specifically, the data decompression device 300 first performs the decoding on the compressed frame, i.e., analyzes the compressed frame using a bitstream analysis module to generate the prediction data PI, the residual data R, and the coded data RAMI. Next, the data decompression device 300 performs prediction based on the prediction data PI to obtain a predicted frame, which is then superimposed with the residual data R to obtain a decoded frame. For ease of explanation, the inventors define the data in the decoded frame as P2. Then, based on the coded data RAMI, the data decompression device 300 performs the edge compensation on the decoded frame P2 using decoded data corresponding to the coded data RAMI to obtain and output the decompressed frame P4. The frame being decompressed may be the decoded frame P2. As described above, the edge compensation may be performed by dividing the decompressed frame (decoded frame P2) into multiple units and using edge compensation coefficients corresponding to the edge adjustment coefficients to perform edge compensation on each unit. The decoded data may include a decoding function and an edge compensation coefficient corresponding to each unit. The decoding function corresponds to the encoding function, and the edge compensation coefficient corresponds to the edge adjustment coefficient. Therefore, the data decompression device 300 may determine the decoding function and the edge compensation coefficient corresponding to each unit based on the encoded data RAMI. Specific details of the decoding function and the edge compensation coefficient corresponding to the edge compensation will be described in detail later. For ease of explanation, the inventors define the transfer function between the decompressed frame P4 and the initial data P0 as the overall spectral adjustment function H0(f). The configuration shown in FIG. 3A reduces the amount of data in the compressed frame, thereby improving the compression rate and encoding efficiency of the initial data, improving transmission efficiency of the initial data, and reducing data loss and avoiding detail loss.

[0064] The data compression device 200 may perform data compression on the initial data by incorporating the edge adjustment into the encoding process. The edge adjustment may be performed at any stage of the encoding process. Accordingly, the edge compensation may also be performed at a corresponding stage of the decoding process.

[0065] FIG. 3B illustrates a flowchart of data compression and data decompression according to an embodiment of the present specification. As shown in FIG. 3B, the data compression device 200 may perform data compression on initial data as follows: The data compression device 200 first predicts the initial frame P0 to obtain a predicted frame and predicted data PI, then performs the edge adjustment on the initial frame P0 and the predicted frame, calculates a residual, obtains the residual data R, and inputs the predicted data PI, the residual data R, and the encoded data RAMI into a bitstream generation module for synthesis to obtain the compressed frame. The frame being compressed may be the predicted frame and the initial frame P0. The specific operation of the data compression illustrated in FIG. 3B is similar to that illustrated in FIG. 3A, except for the order of operations. The edge adjustment will be described in detail later.

[0066] As shown in FIG. 3B , the data decompression device 300 may perform data decompression on the compressed frame as follows: the data decompression device 300 performs the edge compensation while decoding the compressed frame. Specifically, the data decompression device 300 first analyzes the compressed frame using a bitstream analysis module to generate the predicted data PI, the residual data R, and the encoded data RAMI. Then, the data decompression device 300 performs prediction based on the predicted data PI to obtain a predicted frame. Based on the encoded data RAMI, the data decompression device 300 performs the edge compensation on the predicted frame using corresponding decoded data. The edge-compensated predicted frame is superimposed on the residual data R, and the edge compensation is performed on the superimposed data to obtain the decompressed frame P4. The frame being decompressed may be the predicted frame or superimposed data of the predicted frame and the residual data R. For ease of explanation, the inventors define the data in the superimposed frame as P3. Specifically, the specific process of edge compensation for the edges in the superimposed frame P3 will be described in detail later.

[0067] The configuration shown in Figure 3B reduces the amount of data in the compressed frame, thereby improving the compression rate and coding efficiency of the initial data, improving the transmission efficiency of the initial data, and reducing data loss and avoiding loss of details.

[0068] FIG. 3C illustrates a flowchart of data compression and data decompression according to an embodiment of the present specification. As shown in FIG. 3C, the data compression device 200 may perform data compression on initial data as follows: first, the data compression device 200 performs prediction on the initial frame P0 to obtain a residual, thereby obtaining predicted data PI and residual R1; then, the data compression device 200 performs edge adjustment on the residual R1 to obtain the residual data R; and then, the edge-adjusted residual data R, the predicted data PI, and the encoded data RAM1 are input to a bitstream generation module for synthesis to generate the compressed frame. The frame being compressed may be the residual R1. The specific operations of the data compression mode illustrated in FIG. 3C are similar to those illustrated in FIG. 3A, except for the order of operations. The edge adjustment will be described in detail later.

[0069] 3C, the data decompression device 300 may perform data decompression on the compressed frame as follows: the data decompression device 300 analyzes the compressed frame using a bitstream analysis module to generate the prediction data PI and the residual data R, performs prediction based on the prediction data PI to obtain a predicted frame, performs edge compensation on the residual data R using corresponding decoded data based on the encoded data RAMI to obtain a residual R1, and convolves the residual R1 with the predicted frame to obtain the decompressed frame P4. The frame being decompressed may be the residual data R.

[0070] The configuration shown in Figure 3C reduces the amount of data in the compressed frame, thereby improving the compression rate and encoding efficiency of the initial data, improving the transmission efficiency of the initial data, and reducing data loss and avoiding loss of details.

[0071] FIG. 4A shows a flowchart of a data processing method P200 for compressing data. As described above, the data compression device 200 can execute the data processing method P200. Specifically, the storage medium in the data compression device 200 can store at least one instruction set. The instruction set is configured to instruct the compression processor 220 in the data compression device 200 to complete the data processing method P200. When the data compression device 200 operates, the compression processor 220 can read the instruction set and execute the data processing method P200. As shown in FIG. 4A, the method P200 may include:

[0072] S220: An initial frame P0 in the initial data is selected.

[0073] A frame is a processing unit constituting a data sequence. Data processing is often performed frame-by-frame. The initial data may include one or more initial frames. The initial frame P0 includes a predetermined number of bytes of initial data. As described above, this specification uses video data as an example, so the initial data may be initial video data, and the initial frame P0 may be a frame image in the initial video data. In step S220, the data compression device 200 may select some frame images from the initial data as the initial frame P0, or may select all frame images in the initial data as the initial frame P0. The data compression device 200 may select the initial frame P0 according to an application scenario of the initial data. When the initial data is applied to a scenario that does not require high accuracy and compression quality, some frame images may be selected as the initial frame P0. For example, since surveillance images of a remote location are often free of foreign objects, most of the frame images of the surveillance images of the remote location are the same. The data compression device 200 may select some frame images from the initial data as the initial frame P0, compress them, and transmit them. Furthermore, for example, in the case of high-definition television playback video, in order to ensure a good viewing effect, the data compression device 200 may select all frame images as the initial frame P0, compress them, and transmit them.

[0074] S240: The data compression is performed on the initial frame P0 to obtain a compressed frame.

[0075] The data compression may include the edge adjustment and the encoding of the frame being compressed. The edge adjustment of the frame being compressed may involve inputting the frame being compressed into an edge adjustment device and performing edge adjustment. The frame being compressed may include the initial frame P0 and any data state of the initial frame P0 before it becomes the compressed frame in the data compression process. For example, the frame being compressed may include the initial frame P0 and any data state of the initial frame P0 in the process of performing the edge adjustment and encoding, such as an initial frame, a predicted frame, or a residual frame.

[0076] The edge adjustment refers to adjusting the amplitude of the spectrogram of the frame being compressed. For example, the edge adjustment may adjust the amplitude of a selected region within the frequency domain of the frame being compressed, such as the amplitude of a mid-frequency region, the amplitude of a high-frequency region, or further, for example, the amplitude of a low- to mid-frequency region, or further, for example, the amplitude of a mid- to high-frequency region. In some embodiments, the edge adjustment may be achieved by adjusting the amplitude of the selected region within the frequency domain using a single edge adjustment coefficient. For example, the edge adjustment may reduce the amount of data information in the frame being compressed by attenuating the amplitude of the selected region within the frequency domain using an edge adjustment coefficient greater than 0 and less than 1. Because receivers have different frequency sensitivities for different types of data, the data compression operation may select different regions in the frequency domain to attenuate the amplitude depending on the type of data. The mid- to high-frequency components in the data spectrum of one frame are mainly concentrated in regions of rapid data change in the data of that frame, i.e., the edge data of the data. As mentioned above, taking video data as an example, for example, in one frame of an image, the mid- to high-frequency data is mainly concentrated at the edges of objects in the image, i.e., the edge data of the image in this frame. The edges of objects in the image contain a large amount of mid- and high-frequency information, and the mid- and high-frequency regions carry more data. Therefore, reducing the amplitude of the mid- to high-frequency region visually blurs the edge data of the compressed frame and significantly reduces the amount of information in the image. It should be noted that reducing the amplitude of the low-frequency region also reduces the amount of information in the image. In this specification, taking video data as an example, the edge adjustment may be adjusting the amplitude of the mid- to high-frequency region of the compressed frame, for example, attenuating the amplitude of the mid- to high-frequency region to reduce the amount of data information in the mid- to high-frequency region.As will be understood by those skilled in the art, the intermediate state frame that has undergone edge adjustment processing has a higher compression ratio when encoded because the amplitude of the mid- to high-frequency range in the intermediate state frame that has undergone edge adjustment processing is attenuated and the amount of data information is also reduced compared to when the intermediate state frame has not undergone edge adjustment processing.

[0077] Taking video data compression as an example, the data processing method P200 compresses the initial frame P0 using a method that combines edge adjustment and encoding to adjust the amplitude of the mid- to high-frequency region, thereby reducing the amount of data information, further improving the compression rate of video data, and improving the efficiency of video transmission. As mentioned above, when performing data compression on the initial frame P0, the order of the edge adjustment and the encoding may be interchanged or interlaced. Step S240 may include at least one of the data compression methods shown in Figures 3A, 3B, and 3C. For ease of explanation, this specification will describe step S240 in detail using the method shown in Figure 3A as an example. That is, the data compression device 200 first performs edge adjustment on the initial frame P0, attenuating the amplitude in the mid- to high-frequency range of the initial frame P0 to blur the edge information of the initial frame P0, thereby obtaining an encoding-adjusted frame P1. This reduces the amount of information in the initial frame P0, thereby reducing the space resources occupied by the compressed initial frame P0. Next, the encoding-adjusted frame P1 is encoded (i.e., predicted and obtained as a residual), and prediction is performed on the encoding-adjusted frame P1 to obtain a predicted frame and predicted data PI for the encoding-adjusted frame P1. Subsequently, the predicted frame of the encoding-adjusted frame P1 is subtracted from the encoding-adjusted frame P1 to obtain residual data R. The residual data R, the predicted data PI, and the encoded data RAMI are input to a bitstream generation module for synthesis to obtain the compressed frame. The data processing method P200 improves the coding efficiency of the encoding-adjusted frame P1, further reducing the amount of data in the compressed frame, improving coding efficiency, and increasing the compression rate. Since the object of the edge adjustment is the initial frame P0, the frame being compressed is the initial frame P0. Taking video data as an example, in step S240, the data compression of the frame being compressed (initial frame) is performed by at least one compression-side processor 220 of the data compression device 200. S242 may include performing the step of performing the edge adjustment on the frame being compressed (initial frame P0) to obtain the encoding-adjusted frame P1. FIG. 4B shows a flowchart of edge adjustment according to an embodiment of the present specification, and FIG. 5 shows a structural block diagram of edge adjustment according to an embodiment of the present specification. As shown in FIGS. 4B and 5, step S242 may include performing the following steps by at least one compression-side processor 220 of the data compression device 200:

[0078] S242-2: Divide the frame being compressed (initial frame P0) into a plurality of units based on a preset unit size.

[0079] As is known, when encoding image data and video data, the data processing unit may be a frame of data or a portion of a frame of data. For example, a frame of data may be divided into multiple regions, and each region may be coded separately. Taking video data as an example, the region may be a frame or field of image, or a portion of a frame / field of image. For example, in video coding, an image is further divided into slices, tiles, coding units (CUs), macroblocks, blocks, or subblocks. A region is typically an NxN square or an MxN rectangle. The region may have any of the above names, but is not limited to them. For ease of explanation, the inventors define each region as a unit. The edge adjustment can be performed on each unit by treating the unit as an adjustment object. Here, the size of the unit can be arbitrarily selected as necessary, i.e., the values ​​of M and N can be any integer, for example, 4, 8, 16, 32, 16, 128, or 256, or even a smaller value, for example, 2. In some embodiments, the unit may include only one pixel. The higher the resolution of one frame of image, the larger the size of the unit can be. As mentioned above, the frame being compressed (initial frame P0) may be divided into multiple units. For ease of explanation, the inventors will assume that the data of the unit in the ith row and jth column of the initial frame P0 is P0. ij It is defined as:

[0080] S242-4: The amplitude of each unit in the mid- to high-frequency range is adjusted using the corresponding edge adjustment coefficient.

[0081] Taking video data as an example, the edge adjustment may be performed by adjusting the amplitude in the mid- to high-frequency range of each of the units in the frame being compressed (initial frame P0) using the corresponding edge adjustment coefficients, thereby reducing the amplitude in the mid- to high-frequency range of the entire frame being compressed (initial frame P0). The edge adjustment coefficients will be described in detail later. ij When performing the edge adjustment for the current unit P0, step S242-4 ij In contrast, S242-42: Select one function from the preset encoding function group as the encoding function H1 ij (f) and the encoding function H1 ij (f) Current unit P0 ij Adjust the first unit P 1b ij and performing the steps of obtaining a frequency domain such that components in the low frequency region are preserved but components in the mid to high frequency region are attenuated within those frequency regions.

[0082] Here, H1 ij (f) is the current unit P0 in the i-th row and j-th column. ij The encoding function H1 corresponds to ij (f) may be a low-pass filter in the frequency domain, and is a filter for the current unit P0 in the initial frame P0. ij The amplitude of the current unit P0 in the initial frame P0 is smoothly reduced in the frequency domain. ij The current unit P0 ij The first unit P corresponding to 1b ij To reduce the computational complexity and avoid ringing artifacts, we use the coding function H1 ij (f) is the current unit P0 ijThe coding function H1 should smoothly transition the amplitude in the frequency domain. ij (f) may be any form of low pass filter with smooth transition, and this specification is not limited in this regard.

[0083] It should be noted that in order to avoid ringing artifacts, the encoding function H1 ij (f) is a curve with a smooth transition, which avoids abrupt changes in the amplitude adjustment gain in the curve. As mentioned above, the ringing artifacts are caused by the distortion of the selected coding function H1 when performing a spectral adjustment process on an image in image processing. ij (f) refers to the occurrence of "ringing" in the image when there is a rapid change in the image tone. This "ringing" refers to vibrations that occur in areas where the output image has a rapid change in tone, like the air vibrations that occur after a bell is struck. Ringing artifacts often appear at the edges of an image.

[0084] The encoding function H1 ij (f) Current unit P0 ij , adjusting the current unit P0 using a coding convolution kernel in the time domain. ij It may be expressed as performing a convolution on the encoding function H1 ij (f) Current unit P0 ij Adjusting the current unit P0 ij For the transfer function H1 in the frequency domain, ij (f) (i.e., the encoding function), or may be expressed as performing the corresponding convolution calculation in the time domain. ij is the digitized data, the convolution operation is performed by the encoding function H1 ij For ease of explanation, this specification takes the convolution in the time domain as an example, and the encoding function H1 ij (f) Current unit P0 ijHowever, as will be appreciated by those skilled in the art, the coding function H1 ij The form of multiplication with (f) is also within the scope of protection of this specification.

[0085] The encoding function group may be stored in a storage medium of the data compression device 200. The encoding function group may include at least one different encoding function. Each encoding function corresponds to one encoding convolution kernel. That is, the storage medium of the data compression device 200 may include at least one encoding convolution kernel. The data compression device 200 may perform the following steps: ij When performing convolution on the current unit P0, any one function from the encoding function group is used. ij The encoding function H1 corresponds to ij (f), and select the corresponding convolution kernel as the encoding convolution kernel, and select the current unit P0 ij Convolution may be performed on

[0086] When edge adjustment is performed on an image, if the selected encoding function has areas where the values ​​change drastically, a convolution kernel or combination of convolution kernels with a higher order is required. This means an unnecessary increase in the amount of calculations. In addition, a high-order convolution kernel is more likely to cause strong color vibrations, called ringing artifacts, in the output image at locations where the gray level or color changes drastically. Ringing artifacts often appear at image edges. H1 in the frequency domain ij (f) the current unit P0 ij For example, if the low frequency region and the mid frequency region are not connected, the amplitude adjustment gain can be smoothly changed by the encoding function H1. ij (f) is the frequency domain of the current unit P0 ij By adjusting the amplitude of the mid-low frequency region, the change in the amplitude adjustment gain in the mid-low frequency region can be made smooth and continuous.

[0087] The encoding function H1 ij The ratio of the absolute value of the sum of negative coefficients to the sum of non-negative coefficients in the encoding convolution kernel corresponding to (f) is less than 0.1. For example, in some embodiments, all of the convolution kernel coefficients in the encoding convolution kernel may be non-negative. Taking video data as an example, if the encoding convolution kernel contains many negative coefficients, pixel values ​​at image edges will vary significantly. Multiplying large pixel values ​​by negative coefficients will reduce the final convolution result, resulting in darker pixels in the image. If negative numbers appear in the convolution result and the absolute value of the negative numbers is large, calculating the convolution result using unsigned integer arithmetic may result in the inversion of the unsigned integer arithmetic result. Taking the unsigned complement value, which is a negative number, will increase the convolution result, resulting in brighter pixels in the image. Therefore, when designing the coding convolution kernel, all coefficients of the coding convolution kernel may be non-negative, or the ratio of the absolute value of the sum of negative coefficients to the sum of non-negative coefficients in the coding convolution kernel may be less than 0.1, that is, a small number of negative coefficients with small absolute values ​​may be allowed to exist in the coding convolution kernel.

[0088] FIG. 6 illustrates an encoding function H1 according to an embodiment of the present disclosure. ij The horizontal axis is the normalized frequency f, and the vertical axis is the coding function H1 ijThe amplitude adjustment gain H1 of (f) is the normalized frequency f of the horizontal axis. The normalized frequency f of the horizontal axis can be divided into a low frequency region, a low-middle frequency region, a middle frequency region, a middle-high frequency region, and a high frequency region. The definitions of the low frequency, middle frequency, and high frequency regions may differ depending on the type of data. As shown in Figure 6, the maximum value of the normalized frequency of the horizontal axis is 0.5. The high frequency range may include frequencies between (d, 0.5) in the normalized frequency domain, where d is a lower frequency limit of the high frequency range. For example, d may be any one of 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, and 0.45 in the normalized frequency domain. In some embodiments, the high frequency range may include frequencies between (0.33, 0.5) in the normalized frequency domain. For example, the high frequency range may include an interval between any two frequencies among 0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, and 0.5 in the normalized frequency domain, where , 0.5 is the normalized maximum frequency. The mid-frequency range may include frequencies between (b, c), where b is the lower frequency limit of the mid-frequency range and c is the upper frequency limit of the mid-frequency range. For example, the lower frequency limit b of the mid-frequency range may be any one of 0.15, 0.16, 0.17, 0.18, 0.19, 0.2, 0.21, 0.22, 0.23, 0.24, 0.25, 0.26, 0.27, and 0.28 in the normalized frequency domain, and the upper frequency limit c of the mid-frequency range may be any one of 0.35, 0.34, 0.33, 0.32, and 0.31 in the normalized frequency domain. The low-frequency range may include frequencies between [0, a] in the normalized frequency domain, where a is the upper frequency limit of the low-frequency range.The upper frequency limit a of the low frequency region may be any one of 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.10, 0.12, 0.13, 0.14, and 0.15 in the normalized frequency region. If the low frequency region and the mid frequency region are not connected, the frequency region therebetween is called a mid-low frequency region. If the mid frequency region and the high frequency region are not connected, the frequency region therebetween is called a mid-high frequency region.

[0089] Encoding function H1 ij (f) can filter components in the mid- to high-frequency range. ij The stopband interval in (f) may be any interval between frequencies 0.15 and 0.50. For example, the coding function H1 ij The stopband interval in (f) may be within an interval defined by any two of the following values: 0.15, 0.17, 0.19, 0.21, 0.23, 0.25, 0.27, 0.29, 0.31, 0.33, 0.35, 0.37, 0.39, 0.41, 0.43, 0.45, and 0.50. ij The passband section in (f) may be any section between frequencies 0 and 0.35. For example, the encoding function H1 ij The passband interval in (f) may be within an interval defined by any two of the following numerical values: 0, 0.02, 0.04, 0.06, 0.08, 0.10, 0.12, 0.14, 0.15, 0.17, 0.19, 0.21, 0.23, 0.25, 0.27, 0.29, 0.21, 0.23, and 0.35.

[0090] It should be noted that FIG. 6 merely illustrates video data as an example, and as will be appreciated by those skilled in the art, the encoding function H1 ij The curve (f) is not limited to the form shown in FIG. 6, and the current unit P0 ij The amplitude of the current unit P0 in the initial frame P0 is smoothly reduced in the frequency domain. ijAll encoding functions H1 that can maintain the components in the low frequency region in the frequency domain but attenuate the components in the mid to high frequency region ij (f) and the encoding function H1 ij (f) Linear combination H1 ij (f)=Σ m=1 n k m H 1m ij (f), or the encoding function H1 ij (f) product combination H1 ij (f)=Π q=1 n k q H 1q ij (f) or a combination of linear combination and product combination, both of which are within the scope of protection of this specification. ij (f)=Σ m=1 n k m H 1m ij (f) represents a linear combination of n functions, and H1 ij (f) represents the mth function, and k m represents the weight corresponding to the mth function. q≧1 and H1 ij (f)=Π q=1 n k q H 1q ij (f) represents the product combination of n functions, and k q represents the weight corresponding to the qth function, and H 1q ij (f) may be any function.

[0091] Table 1 shows a parameter table of the encoding convolution kernel according to an embodiment of the present specification. Table 1 exemplarily lists the parameters of the encoding convolution kernel, where each row of Table 1 represents one encoding convolution kernel. In the case of an 8-bit video image, the first unit P obtained by the encoding convolution is F ijIn this embodiment, it is necessary to divide the result of the convolution by 16, since it is necessary to ensure that the gradation value of the pixel bit in is within the range of 0 to 255. ij (f) is obtained by Fourier transform based on the current unit P0. Table 1 is merely an example, and those skilled in the art will understand that the encoding convolution kernel is not limited to the parameters shown in Table 1. ij The amplitude of the current unit P0 in the initial frame P0 is smoothly reduced in the frequency domain. ij Any encoding convolution kernel that can preserve low-frequency components in the frequency domain while attenuating mid- to high-frequency components falls within the scope of protection of this specification. [Table 1]

[0092] As described above, in the edge adjustment, the data compression device 200 can execute step S242-42 for each unit. ij is the encoding function H1 ij After being adjusted by (f), the corresponding first unit P 1b ij The encoding function H1 corresponding to all units is obtained. ij The coding function H1(f) is constructed by combining the coding functions H1(f) corresponding to each unit. ij (f) can be combined according to the location of each unit to generate the encoding function H1(f), and the encoding function H1(f) can be regarded as a matrix. It should be noted that when performing step S242-42 for different units, the same encoding function (i.e., encoding convolution kernel) may be selected, or different encoding functions (i.e., encoding convolution kernels) may be selected. That is, different units P0 ij The encoding function H1 corresponds to ij (f) may be the same or different.

[0093] The first unit P corresponds to all units. 1b ij Combine the first frame P 1b The first unit P corresponding to each unit is configured as follows: 1b ij are combined according to the location of each unit to form the first frame P 1b The first frame P 1b is a blurred image. P0 and P 0b The relationship between may be expressed as follows:

number

[0094] As shown in FIG. 4B and FIG. 5, the unit P0 ij When performing the edge adjustment for the current unit P0, step S242-4 ij In contrast, S242-44: Current unit P0 ij and the first unit P 1b ij Calculate the difference between the current unit P0 ij The first edge E corresponding to 1b ij The method may include performing the step of obtaining:

[0095] The mid- to high-frequency components in the data spectrum of each frame are mainly concentrated in the area of ​​the data of this frame where the data changes drastically, i.e., the edge data of the data. For example, in the case of an image of one frame, the mid- to high-frequency data is mainly concentrated at the edges of objects in the image, i.e., the edge data of the image of this frame. The encoding function H1 ij (f) is the current unit P0 ij The amplitude of the first unit P is smoothly reduced in the frequency domain, and the components in the mid- to high-frequency range are attenuated. 1b ij is the current unit P0 ijIt can be understood that the edge information in the current unit P0 ij and the first unit P 1b ij Calculate the difference between the current unit P0 ij edge, i.e., the first edge E 1b ij Therefore, the first edge E 1b ij is the current unit P0 ij The first edge E 1b ij is the current unit P0 ij Each unit P0 ij After step S242-44, the first edges E 1b ij The first edge E corresponding to all units is obtained. 1b ij Combine the first edge frame E 1b The first edge E corresponding to each unit 1b ij The first edge frame E is assembled according to the location of each unit. 1b The first edge frame E 1b may be expressed as follows:

number

[0096] As shown in FIG. 4B and FIG. 5, the unit P0 ij When performing the edge adjustment for the current unit P0, step S242-4 ij In contrast, S242-46: Current unit P0 ij The first edge E is adjusted using the edge adjustment coefficient corresponding to 1b ij Adjust the amplitude of the current unit P0 ij The coding edge E corresponding to 1mij The method may include performing the step of obtaining:

[0097] For ease of explanation, we define the edge adjustment coefficient corresponding to the edge adjustment as g1. ij The edge adjustment coefficient corresponding to g1 ij The image of one frame may have a strong edge, a weak edge, or both a strong edge and a weak edge. The strong edge may be an edge where the difference in pixel values ​​between adjacent pixels is large. The weak edge may be an edge where the difference in pixel values ​​between adjacent pixels is small. In step S242-2, the initial frame P0 is divided into a plurality of units, thereby dividing the initial frame P0 into a plurality of small regions. The smaller the size of the units, the more likely it is that a single edge is included in each unit. If the units are sufficiently small, each unit may include only a strong edge or only a weak edge. Therefore, by dividing the initial frame P0 into a plurality of units in step S242-2, the data compression device 200 can individually adjust the edge information of each unit.

[0098] In the case of image data or video data, encoding the image data or video data using the H.264 / H.265 standard may cause some image and video detail loss. That is, weak edges with small differences between adjacent pixels may have smaller or even disappear after the encoding process, thereby causing detail loss in the image data or video data. Therefore, in order to prevent weak edges with small differences between adjacent pixel bits from being lost during the encoding and decoding process, the edge adjustment enhances the amplitude of weak edges with small differences between adjacent pixel bits in the mid- to high-frequency range, thereby increasing the pixel value difference between adjacent pixels in the weak edges, thereby preventing the weak edges from being lost during encoding, avoiding detail loss, and ensuring that details are still retained after encoding and decoding.

[0099] A strong edge where the difference between adjacent pixels is large will not disappear after the encoding and decoding processes. The edge adjustment attenuates the amplitude of the strong edge where the difference between adjacent pixel bits is large in the medium to high frequency region, thereby reducing the difference in pixel values ​​between adjacent pixels at the strong edge, thereby reducing the data information contained in the strong edge and improving the compression ratio. The strong edge where the difference between adjacent pixel bits is sufficiently large, even if the amplitude is attenuated by the edge adjustment, the retained edge is still sufficiently large and will not disappear after the encoding and decoding processes.

[0100] Taking video data as an example, the edge adjustment coefficient g1 ij The first edge E 1b ij When adjusting the amplitude of the edge adjustment coefficient g1, the information contained in the initial frame P0 should be preserved as much as possible so that the information can be better restored during decompression. ij must be greater than 0. Edge adjustment factor g1 ij The encoded edge E processed by 1mij The amplitude of the middle to high frequency region of is greater than zero, so no data loss occurs. Therefore, all data can be restored when decompressing the compressed data. Otherwise, the edge adjustment coefficient g1 ij If a zero point exists in the unit, data in the mid- to high-frequency range within the unit corresponding to the zero point may be lost, and the decoding side will not be able to restore the lost data during decompression, making it impossible to restore the initial data.

[0101] In step S242-46, the data compression device 200 performs the current unit P0 ij The first edge E corresponding to 1b ij Based on the magnitude of the edge value of the current unit P0 ij The edge adjustment coefficient g1 corresponding to ij can be determined. Different units P0 ij is the different edge adjustment factor g1 ij The edge value may have a first edge E 1b ij Step S242-46 may include:

[0102] S242-462: Current unit P0 ij The first edge E corresponding to 1b ij Determine that the edge value of is less than a first preset threshold value, and set the edge adjustment coefficient g1 greater than 1. ij The first edge E 1b ij Increase the amplitude of

[0103] S242-464: Current unit P0 ij The first edge E corresponding to 1b ij is greater than a second predetermined threshold value; and the edge adjustment coefficient g1 is less than 1. ij The first edge E 1b ij Reduce the amplitude of

[0104] In step S242-462, the current unit P0 ij The first edge E corresponding to 1b ij If the edge value of g1 is smaller than the first threshold, it means that the current unit contains a weak edge with many fine details. In this case, the edge adjustment coefficient g1 is set to be larger than 1. ij Using the current unit P0 ij The first edge E corresponding to 1b ij Adjust the amplitude of the mid- to high-frequency region of the current unit P0 ij The first edge E corresponding to 1b ij The amplitude of the mid- to high-frequency region of the current unit P0 ij maintain a signal-to-noise ratio of 0.01, thereby reducing the current unit P0 ij This prevents weak edges in the image from being lost during data compression (prediction and residual calculation), avoiding detail loss and ensuring coding efficiency. The data processing method P200 improves data compression efficiency and increases the amount of data information for weak edges, thereby avoiding detail loss during the data compression process, i.e., improving data compression efficiency and reducing data distortion. The first threshold may be 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, etc. In some embodiments, the first threshold may be smaller, such as 1, 2, 3, 4, etc. In some embodiments, the first threshold may be any number between 5 and 30.

[0105] In step S242-464, the current unit P0 ij The first edge E corresponding to 1b ij If the edge value at the current unit P0 is greater than the second threshold, ij If g1 contains strong edges, it is set to an edge adjustment coefficient g1 that is smaller than 1 and greater than 0. ij Using the current unit P0ij The first edge E corresponding to 1b ij Adjust the amplitude of the mid- to high-frequency region of the current unit P0 ij The first edge E corresponding to 1b ij By reducing the amplitude of the mid- to high-frequency region of the current unit P0 ij The first edge E corresponding to 1b ij The second threshold may be greater than or equal to the first threshold, thereby reducing the signal strength in the mid- to high-frequency range in the prediction signal, thereby reducing the amount of data information and improving the efficiency of data compression when calculating the residual.

[0106] It should be noted that if the first and second thresholds are not equal, the first edge E 1b ij The edge adjustment may not be performed on the edge values ​​in the range between the first threshold and the second threshold (including the first threshold and the second threshold), and in this case, the edge adjustment coefficient g1 ij may be 1.

[0107] In some embodiments, a first set of edge adjustment coefficients and a second set of edge adjustment coefficients may be pre-stored in the storage medium of the data compression device 200. The first set of edge adjustment coefficients includes at least one coefficient. Each coefficient in the first set of edge adjustment coefficients is greater than 1. The coefficient in the first set of edge adjustment coefficients may be any number greater than 1. When performing step S242-462, the data compression device 200 selects one coefficient from the preset first set of edge adjustment coefficients as the edge adjustment coefficient g1. ij Select the first edge E as 1b ij In some embodiments, the data compression device 200 may select the edge adjustment coefficient g1 from the first preset set of edge adjustment coefficients. ij When selecting the first edge E1b ij The first edge E 1b ij The larger the edge value of g1, the larger the corresponding edge adjustment coefficient g1 ij becomes smaller, and the first edge E 1b ij The smaller the edge value of g1, the greater the corresponding edge adjustment coefficient g1. ij becomes larger.

[0108] The second edge adjustment coefficient group includes at least one coefficient. All of the coefficients in the second edge adjustment coefficient group are greater than 0 and less than 1. The coefficients in the second edge adjustment coefficient group may be any numbers greater than 0 and less than 1. When performing step S242-464, the data compression device 200 selects one coefficient from the preset second edge adjustment coefficient group as the edge adjustment coefficient g1. ij and the amplitude of the first edge E is selected as 1b ij In some embodiments, the data compression device 200 may select the edge adjustment coefficient g1 from the second preset group of edge adjustment coefficients. ij When selecting the first edge E 1b ij The first edge E 1b ij The larger the edge value of g1, the larger the corresponding edge adjustment coefficient g1 ij becomes smaller, and the first edge E 1b ij The smaller the edge value of g1, the greater the corresponding edge adjustment coefficient g1. ij becomes larger.

[0109] In some embodiments, the data compression device 200 further optimizes the current unit P0 by an optimization algorithm. ij The edge adjustment coefficient g1 corresponding to ijSpecifically, the data compression device 200 can establish an optimization equation. For example, the data compression device 200 can determine the current unit P0 ij The optimization target may be the weighted value of the distortion rate and coding rate corresponding to the current unit P0 ij The optimization goal is to minimize the weighted values ​​of the distortion rate and coding rate corresponding to the edge adjustment coefficient g1 ij By performing iterative calculations on the edge adjustment coefficient g1 ij and the edge adjustment coefficient g1 ij The first edge E 1b ij Adjust the amplitude of

[0110] As described above, in the edge adjustment, the data compression device 200 can perform steps S242-S46 for each unit. ij The first edge E corresponding to 1b ij is the edge adjustment coefficient g1 ij After adjusting by 1m ij The edge adjustment coefficient g1 corresponding to all units is obtained. ij The edge adjustment coefficient g1 is formed by combining the above. The edge adjustment coefficient g1 corresponding to each unit ij can be combined according to the location of each unit to generate the edge adjustment coefficient g1. The edge adjustment coefficient g1 can be considered as a matrix. It should be noted that different units P0 ij The edge adjustment coefficient g1 corresponding to ij may be the same or different.

[0111] Encoding edges E corresponding to all units 1m ij Combining the encoded edge frame E 1m The coding edge E corresponding to each unit is constructed. 1m ijare combined according to the location of each unit to form an encoded edge frame E 1m It is possible to generate E 1m and E 1b The relationship between may be expressed as follows:

number

[0112] As shown in FIG. 4B and FIG. 5, the unit P0 ij When performing the edge adjustment for the current unit P0, step S242-4 ij In contrast, S242-48: the first unit P 1b ij and the coding edge E 1m ij By superimposing the current unit P0 ij The coding unit P1 corresponds to ij The method may include performing the step of obtaining:

[0113] Each unit P0 ij The first unit P corresponding to 1b ij and the coding edge E 1m ij After superposition, each of them corresponds to the coding unit P1 ij The coding unit P1 corresponding to all the units is obtained. ij The coding adjustment frame P1 is constructed by combining the coding units P1 corresponding to each unit. ij can be combined according to the location of each unit to generate the coding adjustment frame P1. 1m and P 1b The relationship between may be expressed as follows:

number

[0114] As can be seen from equations (1) to (4), the relationship between P1 and P0 may be expressed by the following equation:

number

[0115] For ease of explanation, we denote the transfer function between P1 and P0 as the encoding transfer function H E (f) is defined, and the coding transfer function H E may be expressed as follows:

number

[0116] In this way, in step S242-2, the initial frame P0 is divided into a plurality of units, thereby dividing the initial frame P0 into a plurality of small regions, and different edge adjustment coefficients g1 are calculated based on the characteristics of each unit. ij It can be seen that by selecting [mathematical formula - see original document], it is possible to more flexibly adjust the component signal strength in the mid- to high-frequency range and achieve a better balance between the coding effect and the bitstream size.

[0117] As shown in FIG. 4A, step S240 may further include: S244: The coding is performed on the coding-adjusted frame P1 (prediction is performed to obtain a residual), and the prediction data PI and the residual data R are obtained. S246: The prediction data PI and the residual data R are input to the bitstream generation module and synthesized to obtain the compressed frame.

[0118] In some embodiments, step S240 includes: S248: Encoding function H1 of each unit in the edge adjustment process of the initial frame P0 ij (f) and edge adjustment coefficient g1 ij may further include inputting the above-mentioned signals to the bitstream generation module and synthesizing them to obtain the compressed frame.

[0119] That is, the coding function H1(f) and edge adjustment coefficient g1 corresponding to the initial frame P0 are input to the bitstream generation module and synthesized to obtain the compressed frame. That is, the compressed frame not only includes the prediction data PI and the residual data R, but also the coding function H1(f) corresponding to each unit of the multiple units. ij (f) and the edge adjustment coefficient g1 ij The encoding function H1 corresponding to each unit is further included. ij (f) and the edge adjustment coefficient g1 ij is the encoded data RAMI. It should be noted that the edge adjustment coefficient g1 is ij That is, one of the plurality of units is used as the basic data, and the edge adjustment coefficient g1 corresponding to the other unit is used as the basic data. ij is the difference between them and the basic data. The edge adjustment coefficient g1 is also ij and a nearby edge adjustment coefficient g1 (i-m,j-n) Here, m and n are both integers and are not simultaneously zero. In this way, the amount of data in the edge adjustment coefficient g1 can be further reduced, and the compression rate can be further improved.

[0120] After the data compression device 200 performs the edge adjustment on the initial frame P0, the coded-adjusted frame P1 is obtained. The amplitude of strong edges from mid- to high-frequency in the coded-adjusted frame P1 is attenuated. The amplitude of weak edges from mid- to high-frequency in the coded-adjusted frame P1 is enhanced. However, by performing the edge adjustment on the initial frame P0, the data compression device 200 reduces the overall data information amount of the coded-adjusted frame P1. The data compression device 200 also performs coding and bitstream generation calculations on the coded-adjusted frame P1 to improve the coding efficiency for the initial frame P0, thereby improving the compression rate of the initial frame and the transmission efficiency of the initial data. At the same time, enhancing the amplitude of weak edges from mid- to high-frequency in the coded-adjusted frame P1 can avoid detail loss.

[0121] As described above, the data processing method P200 can simultaneously perform the edge adjustment on the initial frame, improve the compression rate of the initial frame, improve the coding efficiency and the transmission efficiency of the initial data, reduce data loss, and avoid detail loss.

[0122] FIG. 7A shows a flowchart of a data processing method P300 for decompressing a compressed frame. As described above, the data decompression device 300 can execute the data processing method P300. Specifically, the storage medium in the data decompression device 300 can store at least one instruction set. The instruction set is configured to instruct the decompression processor in the data decompression device 300 to complete the data processing method P300. When the data decompression device 300 operates, the decompression end processor can read the instruction set and execute the data processing method P300. For ease of explanation, the inventors will describe the data processing method P300 in the manner shown in FIG. 3A. The method P300 may include the following:

[0123] S320: Obtain compressed data, which includes the compressed frame.

[0124] The compressed data may include a compressed frame obtained by performing data compression on the initial frame in the initial data using a data processing method P200. The compressed frame includes compressed prediction data PI and residual data R. In some embodiments, the compressed frame is obtained by encoding a coding function H1 corresponding to each unit of a plurality of units in the frame being compressed. ij (f) and edge adjustment coefficient g1 ij , i.e., the encoding function H1(f) and the edge adjustment coefficient g1, i.e., the aforementioned encoded data RAMI. For ease of explanation, in the following description, the inventors will use the example in which the compressed frame further includes the encoding function H1(f) and the edge adjustment coefficient g1. As shown in FIG. 3A, step S320 may include inputting the compressed frame into the bitstream analysis module for analysis and calculation to obtain the predicted data PI and the residual data R, as well as the encoding function H1(f) and the edge adjustment coefficient g1. As mentioned above, in this application, a frame is a general processing unit constituting a data sequence. In data processing, calculations are often performed frame by frame. In the data processing method P200 in which the data compression device 200 compresses data, the initial data may be compressed frame by frame. When the data decompression device 300 decompresses the compressed frame, data decompression may also be performed frame by frame.

[0125] S340: Data decompression is performed on the compressed frame to obtain a decompressed frame.

[0126] The data decompression refers to performing decompression calculations on the compressed frame to obtain a decompressed frame, restoring the decompressed frame to the initial data or substantially restoring the decompressed frame to the initial data, or making the decompressed frame clearer than the initial data. Taking video data as an example, if the amplitude of the decompressed frame at any frequency in the low- to mid-frequency range is restored to or above the threshold of the initial frame, the human eye will find it difficult to distinguish the difference between the decompressed frame and the initial frame. The threshold may be any value between 80% and 90%. For example, the threshold may be any value within a closed interval defined by any two numbers from 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, and 90%. For example, the data decompression should make the amplitude of the decompressed frame at any frequency in the low- to mid-frequency range 85% ±3% or more of the initial frame.

[0127] The data decompression includes edge compensation of the frame being decompressed and decoding the edge-compensated data to obtain the required decompressed frame. The frame being decompressed is one frame of data being decompressed, and includes the compressed frame and any data state of the compressed frame before it becomes the decompressed frame during the decompression process.

[0128] Taking video data as an example, the data processing method P200 compresses the initial frame using a method that combines edge adjustment and encoding to further improve the compression rate of video data and increase the efficiency of video transmission. In video decompression technology, the data processing method P300 can restore data in the compressed frame by using a method that combines decoding (i.e., restoring the compressed frame based on residual data R and prediction data PI) and edge compensation to decompress the compressed frame and obtain a required decompressed frame. The decompressed frame may include the compressed frame and any data state of the compressed frame during the decoding process based on the prediction data PI and residual data R. For example, the decompressed frame may be the compressed frame, a decoded frame obtained by decoding, or a predicted frame obtained by prediction.

[0129] As described above, the data compression may involve adjusting the amplitude of the mid- to high-frequency region of the compressed frame through edge adjustment, for example, attenuating the amplitude of the mid- to high-frequency region to reduce the amount of data information in the compressed frame. Taking video data as an example, the edges of objects in an image contain a large amount of mid- and high-frequency information, and the mid- and high-frequency regions carry more data. Therefore, reducing the amplitude of the mid- to high-frequency region visually blurs the edge data of the compressed frame and significantly reduces the amount of information in the image. Therefore, the data expansion may involve compensating the amplitude of the mid- to high-frequency region of the compressed frame through edge compensation, for example, enhancing the amplitude of the mid- to high-frequency region to restore it to the state of the initial frame or enhance it more than the state of the initial frame.

[0130] The edge compensation applied to the data decompression of the compressed frame refers to inputting the decompressed frame into an edge compensation device to perform edge compensation. The edge compensation may correspond to the edge adjustment, i.e., the edge compensation and the edge adjustment should have a predetermined correlation. By carefully setting the correlation between the edge compensation and the edge adjustment, the edge-adjusted compressed frame, after the edge compensation and data processing, can be completely or substantially restored to the data indicator before the edge adjustment (e.g., image clarity of the image data) without considering other calculation errors, and some indicators can even exceed the data before the encoding adjustment (e.g., the clarity of the decoded image exceeds that of the original image). As described above, the edge adjustment process generates encoded data RAMI (encoding function H1(f)) and edge adjustment coefficient g1. The predetermined correlation between the edge compensation and the edge adjustment may also mean that the decoded data (decoding function H2(f) and edge compensation coefficient g2) in the edge compensation have a predetermined correlation. The correlation between the decoded data (decoding function H2(f) and edge compensation coefficient g2) and the encoded data RAMI (encoding function H1(f) and edge adjustment coefficient g1) and the decoded data (decoding function H2(f) and edge compensation coefficient g2) will be explained in detail later.

[0131] Specifically, step S340 may include: S342: The compressed frame is decoded to obtain a decoded frame P2.

[0132] In the method P300, the frame being decompressed may be the decoded frame P2. The compressed frame may be obtained by encoding the encoding-adjusted frame P1 using the data compression device 200. The data decompression device 300 may decode the compressed frame to obtain the decoded frame P2. That is, a predicted frame is obtained by predicting based on the prediction data PI, and the predicted frame is then superimposed with the residual data R to obtain decoded data P2, which is the data P2 of the decoded frame. There may be some error in the encoding and decoding processes. Assuming that the deviation due to the encoding and decoding processes is very small, the decoded frame P2 and the encoding-adjusted frame P1 are substantially identical, and the relationship between P1 and P2 may be expressed by the following equation:

number

[0133] S344: The edge compensation is performed on the currently decompressed frame (decoded frame P2) to obtain the decompressed frame P4.

[0134] 7B shows a flowchart of edge compensation according to an embodiment of the present disclosure, and FIG. 8 shows a structural flowchart of edge compensation according to an embodiment of the present disclosure. As shown in FIG. 7B and FIG. 8, step S344 may include performing the following steps by at least one decompression edge processor of the data decompression device 300:

[0135] S344-2: Divide the currently decompressed frame (decoded frame P2) into the plurality of units based on the preset unit size. The division method for the currently decompressed frame (decoded frame P2) may be the same as the division method for the currently compressed frame (initial frame P0) in step S242-2, and each unit corresponds to another, and will not be further described here. For ease of explanation, we will divide the data of the unit in the ith row and jth column of the decoded frame P2 into P2. ij It is defined as:

[0136] S344-4: Compensate the amplitude of each unit in the mid- to high-frequency range using an edge compensation coefficient corresponding to the edge adjustment coefficient.

[0137] When performing data decompression on a compressed frame, the data processing method P300 uses the units used during data compression as data decompression units and performs edge compensation on the amplitude of the mid- to high-frequency range of each unit using an edge compensation coefficient corresponding to an edge adjustment coefficient, thereby compensating for the amplitude of the mid- to high-frequency range that was reduced during the data compression process and obtaining a decompressed frame. The edge compensation corresponds to the edge adjustment, and a corresponding relationship exists between the edge compensation coefficient and the edge adjustment coefficient. The edge compensation can restore the edge-adjusted compressed data to the clarity of the initial frame or a clarity higher than that of the initial frame. The edge compensation may be performed by compensating for the amplitude of the mid- to high-frequency range of each of the multiple units of the decompressed frame (decoded frame P2) using an edge compensation coefficient corresponding to the edge adjustment coefficient based on the correlation. The correlation and the edge compensation coefficient will be described in detail later. Unit P2 in the i-th row and j-th column ij When performing the edge compensation for the current unit P2, step S344-4 ij In contrast, S344-42: Decryption function H2 ij (f) is determined, and the decoding function H2 ij (f) Current unit P2 ij Adjust the second unit P 2b ij and performing the steps of obtaining a frequency domain such that components in the low frequency region are preserved but components in the mid to high frequency region are attenuated within those frequency regions.

[0138] Here, H2 ij (f) is the current unit P2 in the i-th row and j-th column. ij represents the decoding function corresponding to the decoding function H2. ij(f) may be a low-pass filter in the frequency domain, and is used to filter the current unit P2 in the decoded frame P2. ij The amplitude of the current unit P2 in the decoded frame P2 is smoothly reduced in the frequency domain. ij The current unit P2 ij The second unit P2 corresponds to ij To reduce the computational complexity required for implementation and to avoid ringing artifacts, we use the decoding function H2 ij (f) is the current unit P2 ij The decoding function H2 should smoothly transition the amplitude in the frequency domain. ij (f) may be any form of low pass filter with smooth transition, and this specification is not limited in this regard.

[0139] It should be noted that in order to avoid ringing artifacts, the decoding function H2 ij (f) is a curve with a smooth transition, avoiding abrupt changes in the amplitude adjustment gain in the curve. As mentioned above, the ringing artifacts are caused by the selected decoding function H2 when performing a spectral adjustment process on an image in image processing. ij (f) refers to the occurrence of "ringing" in the image when there is a rapid change in the image tone. This "ringing" refers to vibrations that occur in areas where the output image has a rapid change in tone, like the air vibrations that occur after a bell is struck. Ringing artifacts often appear at the edges of an image.

[0140] There are various methods for edge compensation processing. In the prior art, edge information may be extracted by directly filtering the decoded frame P2 using a high-pass filter or a band-pass filter, filtering the low-frequency components in the decoded frame P2, and extracting the mid- to high-frequency components in the decoded frame P2. However, the coefficients of the convolution kernels corresponding to the high-pass filter and the band-pass filter tend to have many negative coefficients. As mentioned above, when the convolution kernel has many negative coefficients, strong ringing artifacts may appear in the image obtained by convolution using the convolution kernel. Therefore, to avoid ringing artifacts, the data decompression described in this specification uses a smoothly transitioning decoding function H2 ij (f) to adjust the decoded frame P2, filter the components in the mid-frequency to high-frequency range in the decoded frame P2, and then use the decoded frame P2 and the decoding function H2 ij The edge information can be obtained by calculating the difference with the data adjusted by (f), and the edge information can be adjusted using an edge compensation coefficient to restore it to its initial state or enhance it from its initial state. When obtaining edge information using the above method, the occurrence of ringing artifacts can be avoided by designing a decoding convolution kernel in which all coefficients are non-negative or the ratio of the absolute value of the sum of negative coefficients to the sum of non-negative coefficients is less than 0.1.

[0141] Similar to the edge adjustment, the decoding function H2 ij (f) Current unit P2 ij , by using a decoding convolution kernel in the time domain to adjust the current unit P2 ij Therefore, the decoding convolution kernel used in the edge compensation and the encoding convolution kernel used in the edge adjustment should also have a corresponding correlation. That is, the encoding function H1 ij (f) and the decoding function H2 ij (f) should also have the corresponding correlation. That is, the coding function H1ij (f) and the decoding function H2 corresponding to the encoding convolution kernel ij By selecting (f) and the decoding convolution kernel, the same effect can be achieved in two ways. For ease of explanation, this specification will take the example of performing convolution in the time domain to describe the edge compensation. However, as will be appreciated by those skilled in the art, the decoding function H2 in the frequency domain can be used. ij The present specification also covers the form of performing spectrum adjustment by multiplying with (f).

[0142] When edge adjustment is performed on an image, if the selected decoding function has areas where the values ​​change drastically, a convolution kernel or combination of convolution kernels with a higher order is required to achieve the desired result. This means an unnecessary increase in the amount of calculations. In addition, a high-order convolution kernel is more likely to cause strong color vibrations, called ringing artifacts, in the output image at locations where the gray level or color changes drastically. Ringing artifacts often appear at image edges. H2 in the frequency domain ij (f) the current unit P2 ij For example, if the low frequency region and the mid frequency region are not connected, the amplitude adjustment gain can be smoothly changed by the decoding function H2 ij (f) is the frequency domain of the current unit P2 ij By adjusting the amplitude of the mid-low frequency region, the change in the amplitude adjustment gain in the mid-low frequency region can be made smooth and continuous.

[0143] To avoid ringing artifacts, the ratio of the absolute value of the sum of negative coefficients to the sum of non-negative coefficients in the decoding convolution kernel is less than a threshold. For example, the threshold may be any one of 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1, 0.2, 0.3, and 0.4, or any value within an interval defined by any two numbers. For example, all convolution kernel coefficients in the decoding convolution kernel may be selected to be non-negative. Decoding function H2 ij In (f), the amplitude adjustment gain for the mid- to high-frequency range is 0 and can vary within a certain error range. The error range may be within an interval defined by any two of the following values: 0, ±1%, ±2%, ±3%, ±4%, ±5%, ±6%, ±7%, ±8%, ±9%, ±10%, ±11%, ±12%, ±13%, ±14%, ±15%, ±16%, ±17%, ±18%, ±19%, ±20%, ±21%, ±22%, ±23%, ±24%, ±25%, ±26%, ±27%, ±28%, ±29%, ±30%, ±31%, ±32%, ±33%, ±34%, ±35%, etc.

[0144] Decryption function H2 ij (f) ensures that the amplitude adjustment gain is 1 in the DC portion, i.e., the portion where the frequency is 0, and that the basic information in the initial frame can be maintained. ij (f) Current unit P2 ij When adjusting the amplitude adjustment gain, the amplitude adjustment gain for the low frequency region smoothly transitions from an amplitude adjustment gain of 1 where the frequency is 0 to an amplitude adjustment gain close to 0 in the mid frequency region.

[0145] As mentioned above, the encoding function H1 ij (f) and the decoding function H2 ij (f) should also have the corresponding correlation. That is, the decoding function H2 ij (f) is the encoding function H1 ij(f). A set of decoding functions may be stored in the storage medium of the data decompression device 300. The set of decoding functions may include at least one different decoding function. Each decoding function corresponds to one decoding convolution kernel. That is, the storage medium of the data decompression device 300 may include at least one decoding convolution kernel. In some embodiments, the storage medium of the data decompression device 300 may include a set of decoding functions, such as H2 ij (f) and encoding function H1 ij The correspondence between (f) may also be stored.

[0146] As mentioned above, in some embodiments, the compressed frame includes the coded data RAMI, i.e., the coding function H1(f) and the edge adjustment coefficient g1, i.e., the coding function H1 corresponding to each unit. ij (f) and edge adjustment coefficient g1 ij The data decompression device 300 includes the current unit P2 ij When performing convolution on ij (f) and encoding function H1 ij (f) from the preset decoding function group to the current unit P2 ij The encoding function H1 ij (f) The function corresponding to the current unit P2 ij The decoding function H2 corresponding to ij (f) selecting a current unit P2 from the at least one decoding convolution kernel; ij The decoding function H2 corresponding to ij (f) is selected as the decoding convolution kernel, and the current unit P2 ij Convolution can be performed on In some embodiments, the compressed frame does not include the coded data RAMI, i.e., the coding function H1(f) and the edge adjustment coefficient g1. In this case, the data decompression device 300 decodes the current unit P2 ijWhen performing convolution on the decoding function H2, one function is arbitrarily selected from the preset decoding function group. ij (f), and the corresponding convolution kernel is used as the decoding convolution kernel. ij The data decompression device 300 also selects a decoding function H2 from the group of decoding functions depending on empirical values. ij (f) is selected, for example, by machine learning to generate the decoding function H2 ij (f) can be selected.

[0147] In some embodiments, the decoding function H2 ij (f) is the encoding function H1 ij (f). The graph shown in Figure 6 shows the decoding function H2 ij As will be appreciated by those skilled in the art, the current unit P2 ij The amplitude of the current unit P2 in the decoded frame P2 is smoothly reduced in the frequency domain. ij All decoding functions H2 that can preserve the components in the low frequency region in the frequency domain but attenuate the components in the mid to high frequency region ij (f) and the decoding function H2 ij Linear combination H2 of (f) ij (f)=Σ m=1 n k m H 2m ij (f), or the decoding function H2 ij (f) product combination H2 ij (f)=Π q=1 n k q H 2q ij (f) or a combination of linear combination and product combination, both of which are within the scope of protection of this specification. ij (f)=Σ m=1 n k m H 2m ij (f) represents a linear combination of n functions, and H2m ij (f) represents the mth function, and k m represents the weight corresponding to the mth function. q ≥ 1, and H2 ij (f)=Π q=1 n k q H 2q ij (f) represents the product combination of n functions, and k q represents the weight corresponding to the qth function, and H 2q ij (f) may be any function.

[0148] As described above, in the edge compensation, the data decompression device 300 can execute step S344-42 for each unit. ij is the decoding function H2 ij After being adjusted by (f), the corresponding second unit P 2b ij The decoding function H2 corresponding to all units is obtained. ij (f) to obtain the decoding function H2(f). The decoding function H2 corresponding to each unit is ij (f) can be combined according to the location of each unit to generate the decoding function H2(f), which can be considered as a matrix.

[0149] A second unit P that corresponds to all units 2b ij Combine the two to create the second frame P 2b The second unit P corresponding to each unit is configured. 2b ij The units are combined according to their locations to form the second frame P 2b The second frame P 2b is a blurred image. P2 and P 2b The relationship between may be expressed as follows:

number

[0150] As shown in FIG. 7B and FIG. 8, the unit P2 ij When performing the edge compensation for the current unit P2, step S344-4 ij In contrast, S344-44: Current unit P2 ij and the second unit P 2b ij Calculate the difference between the current unit P2 ij The second edge E corresponding to 2b ij The method may further include performing the step of obtaining:

[0151] The mid- to high-frequency components in the data spectrum of each frame are mainly concentrated in the area of ​​the data of this frame where the data changes drastically, i.e., the edge data of the data. For example, in the case of an image of one frame, the mid- to high-frequency data is mainly concentrated at the edges of the objects in the image, i.e., the edge data of the image of this frame. The decoding function H2 ij (f) is the current unit P2 ij The amplitude of the second unit P is smoothly reduced in the frequency domain, and the components in the mid- to high-frequency range are attenuated. 2b ij is the current unit P2 ij It can be understood that the edge information in the current unit P2 is removed. ij and the second unit P 2b ij Calculate the difference between the current unit P2 ij edge, i.e., the second edge E 2b ij Therefore, the second edge E 2b ij is the current unit P2 ij The second edge E 2b ij is the current unit P2 ij Each unit P2 contains edge information. ijAfter step S344-44, the corresponding second edge E 2b ij The second edge E corresponding to all units is obtained. 2b ij Combine with the second edge frame E 2b The second edge E corresponding to each unit is 2b ij The second edge frame E is assembled according to the location of each unit. 2b The second edge frame E 2b may be expressed as follows:

number

[0152] As shown in FIG. 7B and FIG. 8, the unit P2 ij When performing the edge compensation for the current unit P2, step S344-4 ij In contrast, S344-46: Current unit P2 ij Edge adjustment coefficient g1 ij The second edge E is calculated using the edge compensation coefficient corresponding to 2b ij Compensate for the amplitude of the current unit P2 ij The decoded edge E corresponding to 2m ij The method may further include performing the step of obtaining:

[0153] For ease of explanation, we define the edge compensation coefficient corresponding to the edge compensation as g2. We define the unit P2 in the i-th row and j-th column as ij The edge compensation coefficient corresponding to g2 ijAs described above, the edge adjustment attenuates the amplitude of a strong edge in the frame being compressed in the mid- to high-frequency region of that frequency region, and blurs the edge data of the frame being compressed, thereby reducing the amount of data required for encoding. The edge compensation can restore and further enhance the data to which the edge adjustment has been applied. In other words, the edge compensation can completely or substantially restore the amplitude of the mid- to high-frequency region of the frame being expanded to the state before attenuation, and can further enhance it compared to the state before attenuation. In other words, the edge adjustment coefficient g1 ij and edge compensation coefficient g2 ij and have the previously set correlation, that is, the edge adjustment coefficient g1 ij and edge compensation coefficient g2 ij correspond to each other. A set of edge compensation coefficients may be stored in the storage medium of the data decompression device 300. The set of edge compensation coefficients may include at least one different coefficient. In some embodiments, the storage medium of the data decompression device 300 stores edge adjustment coefficients g1, ij and edge compensation coefficient g2 ij A correspondence relationship between the

[0154] As mentioned above, in some embodiments, the compressed frame includes the coded data RAMI, i.e., the coding function H1(f) and the edge adjustment coefficient g1, i.e., the coding function H1 corresponding to each unit. ij (f) and edge adjustment coefficient g1 ij When performing step S344-46, the data decompression device 300 uses the edge adjustment coefficient g1 ij and edge compensation coefficient g2 ij Based on the correlation with the current unit P2, the current unit P2 is selected from the preset edge compensation coefficient group. ij Edge adjustment coefficient g1 ij The edge compensation coefficient g2 corresponding to ij The current unit P2 ij The edge compensation coefficient g2 corresponding to ij and select the second edge E as 2b ijThe amplitude of the signal can be compensated for.

[0155] In some embodiments, the compressed frame does not include the encoded data RAMI, i.e., the encoding function H1(f) and the edge adjustment coefficient g1. In this case, when performing step S344-46, the data decompression device 300 arbitrarily selects one coefficient from the preset set of edge compensation coefficients as the current unit P2. ij The edge compensation coefficient g2 corresponding to ij and select the second edge E as 2b ij The data decompression device 300 can also select an edge compensation coefficient g2 from the group of edge compensation coefficients based on empirical values. ij Select, for example, the edge compensation coefficient g2 by machine learning. ij You can select:

[0156] Decoding edge E corresponding to all units 2m ij Decode the edge frame E by combining 2m The decoding edge E corresponding to each unit is constructed. 2m ij The edge frame E is decoded by combining the units according to their positions. 2m It is possible to generate E 2m and E 2b The relationship between may be expressed as follows:

number

[0157] As shown in FIG. 7B and FIG. 8, the unit P0 ij When performing the edge compensation for the current unit P2, step S344-4 ij In contrast, S344-48: Current unit P2 ij and the decoded edge E 2m ij By superimposing the current unit P2 ij The decoding unit P4 corresponds toij The method may include performing the step of obtaining:

[0158] Each unit P2 ij and the decoded edge E 2m ij After superposition, each of them is decoded by the corresponding decoding unit P4 ij The decoding unit P4 corresponding to all the units is obtained. ij The decompressed frame P4 is constructed by combining the above units. ij can be combined according to the location of each unit to generate the decompressed frame P4. 2m The relationship between P and P2 may be expressed as follows:

number

[0159] As can be seen from equations (8) to (11), the relationship between P2 and P4 may be expressed by the following equation:

number

[0160] For ease of explanation, we will denote the transfer function between P4 and P2 as the decoded transfer function H D (f) and the decoding transfer function H D may be expressed as follows:

number

[0161] Based on equations (5), (6), (12) and (13), the relationship between P0 and P4 may be expressed as follows:

number

[0162] Taking video data as an example, because the human eye is more sensitive to information in the low- to mid-frequency range, the edge adjustment design attenuates only the amplitude of the mid- to high-frequency range in the initial frame P0, and the low-frequency information of the initial frame P0 is retained in the coding-adjusted frame P1. Because the decoded frame P2 and the coding-adjusted frame P1 are substantially identical, the low-frequency information is also retained in the decoded frame P2. Meanwhile, the edge compensation design compensates only the amplitude of the mid- to high-frequency range in the decoded frame P2. Therefore, the low-frequency information of the initial frame P0 is retained in the decompressed frame P4. Theoretically, without considering deviations due to other algorithms, the decompressed frame P4 obtained by compensating the amplitude of the mid- to high-frequency range in the decoded frame P2 using the edge compensation can completely or substantially restore all of the mid-frequency information of the initial frame P0. That is, the data decompression can restore and even enhance the compressed data at any mid-frequency range. Therefore, after data decompression, the amplitude of the decompressed frame P4 at any frequency in the low frequency range must be approximately equal to that of the initial frame P0, and the amplitude of any frequency in the mid-frequency range must be approximately equal to or greater than that of the initial frame P0. The term "approximately equal" refers to the amplitude of the decompressed frame P4 being equal to that of the initial frame P0, fluctuating within a certain error range. Taking video data as an example, if the amplitude of the decompressed frame P4 at any frequency in the low-to-mid frequency range is restored to 85% or more of that of the initial frame P0, the human eye will find it difficult to distinguish the difference between the decompressed frame P4 and the initial frame P0. Therefore, after data decompression, the amplitude of the decompressed frame P4 at any frequency in the low-to-mid frequency range must be 85% or more of that of the initial frame P0. That is, the error range should not be such that the amplitude of the decompressed frame P4 at any frequency in the low-to-mid frequency range is less than 85% of that of the initial frame P0.Since the human eye is not sensitive to high-frequency information, the information in the high-frequency region of the decompressed frame P4 may be retained to accommodate scenarios requiring high quality, or may be attenuated to suppress unwanted high-frequency noise. That is, the amplitude of any high-frequency frequency of the decompressed frame P4 may be approximately equal to, lower than, or even higher than that of the initial frame P0. The relationship between P0 and P4 may be expressed by the following equation:

number

number

[0163] It should be noted that the formula allows for a certain range of error. For example, P4 ≧ P0 may mean that when the reference value of P4 is equal to or greater than P0, P4 is allowed to fluctuate within a certain error range. In other words, when P4 = P0, P4 is allowed to be slightly smaller than P0 in the case of a negative error. The formula here merely shows the basic relationship between P4 and P0, and does not include any error. As those skilled in the art will understand, even if the amplitude of the low- to mid-frequency range of the expanded frame P4 is slightly smaller than that of the initial frame P0 due to fluctuations within the error range, this specification also covers the case where this occurs. The following formula also allows for a certain range of error. In the following, only the basic relationship that the amplitude of P4 is equal to or greater than the initial frame P0 will be described. Those skilled in the art will be able to deduce the fluctuations within the error range themselves.

[0164] For ease of explanation, if we define the overall spectral adjustment function between P0 and P4 as H0(f), the relationship between P0 and P4 may be expressed as:

number

[0165] If so, the overall spectral adjustment function H0(f) may be expressed as:

number

number

[0166] In equations (18) and (19), when H0(f)≈1, data within the selected frequency domain section of the decompressed frame P4 can be restored to the initial frame P0. When H0(f)>1, data within the selected frequency domain section of the decompressed frame P4 can be enhanced, i.e., the amplitude of the selected region of the decompressed frame P4 is higher than that of the initial frame P0. For example, if the initial frame P0 is a frame in a video, increasing H0(f) to greater than 1 in the selected frequency domain section can achieve intelligibility enhancement. For ease of explanation, we define H0(f)≈1 as the normal mode and H0(f)>1 as the enhancement mode. Hereinafter, we will use video data as an example to explain the global spectral adjustment function H0(f) in detail.

[0167] FIG. 9A shows a graph of the overall adjustment function H0(f) according to an embodiment of the present specification. FIG. 9B shows a graph of the overall adjustment function H0(f) according to an embodiment of the present specification. FIG. 9C shows a graph of the overall adjustment function H0(f) according to an embodiment of the present specification. FIG. 9D shows a graph of the overall adjustment function H0(f) according to an embodiment of the present specification. As shown in FIGS. 9A to 9D, the horizontal axis represents the normalized frequency f, and the vertical axis represents the amplitude adjustment gain H0 of the overall spectral adjustment function H0(f). The graphs in FIGS. 9A to 9D represent different overall spectral adjustment functions H0(f). The maximum value of the normalized frequency on the horizontal axis is 0.5. The normalized frequency f on the horizontal axis can be divided into a low frequency region, a mid-low frequency region, a mid-frequency region, a mid-high frequency region, and a high frequency region. The frequencies between (0, a) belong to the low frequency range, the frequencies between (a, b) belong to the low-medium frequency range, the frequencies between (b, c) belong to the medium frequency range, the frequencies between (c, d) belong to the high-medium frequency range, and the frequencies between (d, 0.5) belong to the high frequency range. Here, the values ​​of a, b, c, d, and e refer to FIG. 6 and will not be further described here.

[0168] Because the human eye is more sensitive to low- to mid-frequency data than high-frequency data in video data, after data decompression, the information in the low- to mid-frequency region of the decompressed frame P4 should be maintained relative to the initial frame P0, and not lost. That is, the global spectral adjustment function H0(f) should make the amplitude of the low- to mid-frequency region of the decompressed frame P4 at least 85% of that of the initial frame P0, and may even be greater than that of the initial frame P0. Because the human eye is not sensitive to information in the high-frequency region, the amplitude of the high-frequency region of the decompressed frame P4 can be selected according to different application scenarios. For example, in a scenario with low clarity requirements, the amplitude of the high-frequency region of the decompressed frame P4 may be smaller than that of the initial frame P0. In a reconnaissance scenario with very high clarity requirements, the amplitude of the high-frequency region of the decompressed frame P4 may be approximately equal to or greater than that of the initial frame P0. 9A-9D, when the amplitude adjustment gain H0 of the global adjustment function H0(f) at any frequency f in the low- to mid-frequency range (including the low- and mid-frequency ranges) is greater than or approximately equal to 1, the amplitude of the expanded frame P4 after expansion is less than 85% of that of the initial frame P0, restoring or enhancing clarity and improving the visual observation effect. Here, the aforementioned "approximately equal to 1" may refer to fluctuations within a certain error range equal to 1. The error range may be within an interval defined by any two of the following values: 0, ±1%, ±2%, ±3%, ±4%, ±5%, ±6%, ±7%, ±8%, ±9%, ±10%, ±11%, ±12%, ±13%, ±14%, ±15%, etc. For ease of explanation, the inventors define the amplitude adjustment gain in the high frequency region of the overall adjustment function H0(f) as the first amplitude adjustment gain, the amplitude adjustment gain in the mid frequency region as the second amplitude adjustment gain, and the amplitude adjustment gain in the low frequency region as the third amplitude adjustment gain. The third amplitude adjustment gain value, the second amplitude adjustment gain value, and the first amplitude adjustment gain value can vary within the error range.

[0169] As shown in FIG. 9A, the third amplitude adjustment gain value, the second amplitude adjustment gain value, and the first amplitude adjustment gain value in the low-frequency to high-frequency region of the overall adjustment function H0(f) are all approximately equal to 1, so that the amplitudes in the low-frequency to high-frequency region of the extended frame P4 are all 85% or more of those of the initial frame P0, so that the data in the low-frequency to high-frequency region of the extended frame P4 can be smoothly restored or substantially restored to the state of the initial frame P0.

[0170] As shown in FIG. 9B , the third amplitude adjustment gain value and the second amplitude adjustment gain value in the low- to mid-frequency region of the global adjustment function H0(f) are both approximately equal to 1, thereby gently restoring or substantially restoring the data in the low- to mid-frequency region of the decompressed frame P4 to the state of the initial frame P0. The first amplitude adjustment gain value in the high-frequency region of the global adjustment function H0(f) is less than 1, thereby gently reducing the amplitude in the high-frequency region of the decompressed frame P4 relative to the initial frame P0 to suppress high-frequency noise. The gentle reduction in amplitude may be achieved by attenuating the amplitude by the first amplitude adjustment gain value or by attenuating the amplitude within a certain error range around the first amplitude adjustment gain value. For example, the first amplitude adjustment gain may be any value between 0 and 1. For example, the first amplitude adjustment gain value may be within an interval defined by any two of the following values: 0, 0.04, 0.08, 0.12, 0.16, 0.20, 0.24, 0.28, 0.32, 0.36, 0.40, 0.44, 0.48, 0.52, 0.56, 0.60, 0.64, 0.68, 0.72, 0.76, 0.80, 0.84, 0.88, 0.92, 0.96, and 1. As shown in FIG. 9B, the first amplitude adjustment gain in the high-frequency region (approximately the interval between 0.4 and 0.5) of the overall adjustment function H0(f) is approximately 0.6. The second and third amplitude adjustment gain values ​​are both near 1. The second and third amplitude adjustment gain values ​​can vary within a certain error range, for example, the second and third amplitude adjustment gain values ​​may be within an interval defined by any two of numerical values ​​such as 0.85, 0.90, 0.95, 1, 1.05, 1.10, and 1.15.

[0171] 9C , the third amplitude adjustment gain value in the low-frequency region of the global adjustment function H0(f) is approximately equal to 1, thereby gently restoring or substantially restoring the data in the low-frequency region of the expanded frame P4 to the state of the initial frame P0. The second amplitude adjustment gain value in the mid-frequency region and the first amplitude adjustment gain value in the high-frequency region of the global adjustment function H0(f) are both greater than 1, thereby gently increasing the amplitude in the mid- to high-frequency region of the expanded frame P4 relative to the initial frame P0, thereby enhancing the clarity of the data in the mid- to high-frequency region. The gentle increase in amplitude may be such that the amplitude is increased by the second amplitude adjustment gain value and the first amplitude adjustment gain value, or may be such that the amplitude is increased within a certain error range around the second amplitude adjustment gain value and the first amplitude adjustment gain value. The second amplitude adjustment gain value and the first amplitude adjustment gain value may be substantially equal in magnitude, or the second amplitude adjustment gain value may be larger than the first amplitude adjustment gain value, or the second amplitude adjustment gain value may be smaller than the first amplitude adjustment gain value. In the graph shown in FIG. 9C , the second amplitude adjustment gain value and the first amplitude adjustment gain value are substantially equal in size. The second amplitude adjustment gain value and the first amplitude adjustment gain value may be any numerical value greater than 1. For example, the second amplitude adjustment gain value and the first amplitude adjustment gain value may be within an interval defined by any two of numerical values ​​such as 1, 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 2, 2.1, 2.2, and 2.4. As shown in FIG. 9C , the second amplitude adjustment gain and the first amplitude adjustment gain of the global adjustment function H0(f) in the mid- to high-frequency region are approximately 1.2.

[0172] As shown in FIG. 9D, the third amplitude adjustment gain value in the low-frequency region of the global adjustment function H0(f) is approximately equal to 1, thereby gently restoring or substantially restoring the data in the low-frequency region of the expanded frame P4 to the state of the initial frame P0. The second amplitude adjustment gain value in the mid-frequency region of the global adjustment function H0(f) is greater than 1, thereby gently increasing the amplitude in the mid-frequency region of the expanded frame P4 relative to the initial frame P0, thereby enhancing data clarity in the mid-frequency region. The first amplitude adjustment gain value in the high-frequency region of the global adjustment function H0(f) is less than 1, thereby gently decreasing the amplitude in the high-frequency region of the expanded frame P4 relative to the initial frame P0, thereby reducing the amount of data in the non-sensitive high-frequency region and suppressing high-frequency noise. The graph shown in FIG. 9D can reduce the amount of data while enhancing clarity. The second amplitude adjustment gain value may be any value greater than 1. The first amplitude adjustment gain may be any value between 0 and 1. As shown in FIG. 9D, the overall adjustment function H0(f) has a second amplitude adjustment gain of about 1.2 in the mid-frequency region and a first amplitude adjustment gain of about 0.6 in the high-frequency region.

[0173] Furthermore, if the high-frequency region and the mid-frequency region are not connected, the overall spectral adjustment function H0(f) can further adjust the amplitude of the high-frequency region to make the change of the amplitude adjustment gain in the mid-to-high frequency region smooth and continuous.

[0174] Furthermore, if the mid-frequency region and the low-frequency region are not connected, the overall spectral adjustment function H0(f) can further adjust the amplitude of the mid-low frequency region to make the change of the amplitude adjustment gain continuous in the mid-low frequency region.

[0175] The curve of the global adjustment function H0(f) is a smoothly transitioning curve. In engineering practice, the curve of the global adjustment function H0(f) may be allowed to have a small range of fluctuations that do not affect the effect of decompression, provided that the amplitude in the low- to mid-frequency range of the decompressed frame P4 is approximately equal to or greater than that of the initial frame P0. For data in other forms than video data, the parameters of the global adjustment function H0(f) may be set based on the receiver's sensitivity to the data. The receiver's sensitivity to the data varies depending on the form of the data.

[0176] For ease of explanation, we take the situation shown in equation (18) as an example. Combined with equations (17) and (18), the decompressed frame P4 may be expressed as follows:

number

[0177] FIG. 10A shows the overall adjustment function H0(f), the coding function H1(f), and the coding transfer function H E (f) and the decoding transfer function H D 10B shows a graph of the overall adjustment function H0(f), the coding function H1(f), and the coding transfer function H(f) of the enhancement mode according to an embodiment of the present disclosure. E (f) and the decoding transfer function H D10A and 10B show graphs of the normalized frequency f and amplitude adjustment gain H. The encoding and decoding convolution kernels used in FIGS. 10A and 10B are the same, and are the convolution kernels shown in Table 1. In FIG. 10A, the edge adjustment coefficient g1 is 0.5, and the edge compensation coefficient g2 is 0.96. In FIG. 10B, the edge adjustment coefficient g1 is 0.5, and the edge compensation coefficient g2 is 1.6. As shown in FIGS. 10A and 10B, the horizontal axis represents the normalized frequency f, and the vertical axis represents the amplitude adjustment gain H. As shown in FIG. 10A, the global spectral adjustment function H0(f) is approximately 1 in any frequency domain. The global spectral adjustment function H0(f) performs normal-mode spectral adjustment on the decompressed frame. That is, the global spectral adjustment function H0(f) completely preserves all frequency information, and the data in the decompressed frame can be substantially restored to the data in the initial frame. As shown in Figure 10B, the global spectral adjustment function H0(f) in the low frequency region is ≈1, and the global spectral adjustment function H0(f) in the mid- to high frequency region is >1. The global spectral adjustment function H0(f) performs an enhancement mode spectral adjustment on the mid- to high frequency region of the decompressed frame, i.e., the global spectral adjustment function H0(f) enhances information related to the mid- to high frequency region, and the data in the mid- to high frequency region in the decompressed frame is enhanced compared to the data in the mid- to high frequency region in the initial frame. It should be noted that the graphs shown in Figures 10A and 10B are merely illustrative, and those skilled in the art will appreciate that the global spectral adjustment function H0(f), H1(f), and H E (f) and H D The graph of (f) is not limited to the form shown in FIGS. 10A and 10B, and may be any of H0(f), H1(f), H2(f), and H3(f) that satisfy the formulas (18) and (19). E (f) and H D (f) All graphs are within the scope of protection of this specification.

[0178] As described above, when compressing the initial data, the data processing system 100 according to this specification uses the data compression device 200 to execute the method P200. This method divides an initial frame in the initial video data into multiple units, obtains the amplitude of the mid- to high-frequency region of each unit, and adjusts the amplitude of the mid- to high-frequency region of each unit using different edge adjustment coefficients to reduce the amplitude of the mid- to high-frequency region of the initial frame. If the amplitude of the mid- to high-frequency region of the current unit is large, indicating that the current unit contains a strong edge, the edge adjustment coefficient is less than 1 and greater than 0 to adjust the amplitude of the mid- to high-frequency region of the current unit, thereby reducing the signal strength in the mid- to high-frequency region of the current unit, thereby reducing the amount of data information and improving the efficiency of data compression when predicting and calculating the residual. If the amplitude of the mid-frequency to high-frequency region of the current unit is small, indicating that the current unit contains a weak edge, an edge adjustment coefficient greater than 1 is used to adjust the amplitude of the mid-frequency to high-frequency region of the current unit, thereby increasing the amplitude of the mid-frequency to high-frequency region of the current unit, thereby preventing the weak edge of the current unit from being lost during data compression (prediction and residual determination) and avoiding detail loss.The data processing method P200 and system 100 improve data compression efficiency and increase the data information content of weak edges, thereby preventing detail loss during data compression, i.e., improving data compression efficiency and reducing data distortion.

[0179] In the data processing system 100 according to this specification, when data decompression is performed on the compressed frame, the data decompression device 300 executes the method P300. Using the data compression units as data decompression units, edge compensation is performed on the amplitude of the mid- to high-frequency range of each unit using edge compensation coefficients corresponding to the edge adjustment coefficients. This compensates for the amplitude loss in the mid- to high-frequency range during the data compression process, thereby obtaining a decompressed frame. The edge compensation corresponds to the edge adjustment, and there is a corresponding relationship between the edge compensation coefficients and the edge adjustment coefficients. The edge compensation can restore the edge-adjusted compressed data to the clarity of the initial frame or a clarity higher than that of the initial frame. In other words, the decoding side can restore data within at least the important frequencies of the decompressed data to the clarity of the initial frame without significantly increasing the amount of encoding and decoding calculations, and can even obtain a clarity higher than that of the initial frame. Since the edge adjustment coefficients in the edge adjustment process of the initial frame are all greater than 0 and no information is lost in the compressed frame, the edge adjustment coefficients and edge compensation coefficients can be designed based on the relationship between the edge adjustment coefficients and the edge compensation coefficients and their respective characteristics to restore the information in the compressed frame.The method and system can significantly improve data compression efficiency, improve data transmission efficiency, reduce data loss, avoid detail loss, remove noise, and improve the clarity of the decompressed data. The present specification also provides a non-transitory storage medium having stored thereon at least one set of executable instructions for data processing, which, when executed by a processor, instruct the processor to perform steps of a data processing method P200. In some possible embodiments, each aspect of the present specification may be realized in the form of a program product including program code. When the program product runs on the data compression device 200, the program code causes the data compression device 200 to perform the data processing steps described herein. A program product for implementing the above method may be a portable compact disc read-only memory (CD-ROM), include the program code, and be operable on the data compression device 200, e.g., a personal computer. However, the program product of the present specification is not limited thereto. In the present specification, a readable storage medium may be any tangible medium that contains or stores a program, and the program may be used by or in combination with an instruction execution system (e.g., the compression-side processor 220). The program product may use any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. Further examples of the computer-readable storage medium include an electrical connection having one or more conductors, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), optical fiber, a portable compact disk read-only memory (CD-ROM), an optical memory device, a magnetic memory device, or any suitable combination thereof. The computer-readable storage medium may include a data signal propagated in baseband or as part of a carrier wave, in which the computer-readable program code is embedded.Such propagated data signals may take various forms, such as, but not limited to, electromagnetic signals, optical signals, or any suitable combination thereof. A readable storage medium may also be any readable medium other than a readable storage medium, which may transmit, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained in the readable storage medium may be transmitted over any suitable medium, including, but not limited to, wireless, wired, optical cable, RF, or the like, or any suitable combination of the above. The program code for performing the operations herein may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, or the like, general procedural programming languages ​​such as "C," or similar programming languages. The program code may be executed entirely by the data compression device 200, partially by the data compression device 200, as separate software packets, partially by the data compression device 200 and partially by a remote computing device, or entirely by a remote computing device. In the case of a remote computing device, the remote computing device may be connected to the data compression apparatus 200 via the transmission medium 120 or may be connected to an external computing device.

[0180] The foregoing describes specific examples of the present specification. Other examples are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the examples and still achieve desirable results. Also, processes depicted in the figures do not necessarily require a particular order or sequential order to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or advantageous.

[0181] As such, those skilled in the art will appreciate upon reading this detailed disclosure that the foregoing detailed disclosure may be presented by way of example only and not by way of limitation. Although not expressly stated herein, this specification should cover various reasonable changes, improvements, and modifications to the examples, as would be understood by those skilled in the art. These changes, improvements, and modifications are intended to be presented by this specification and are within the spirit and scope of the illustrative examples herein.

[0182] It should be noted that certain terms in this specification are used to describe embodiments of this specification. For example, "one embodiment," "embodiment," and / or "some embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of this specification. Therefore, it should be emphasized and understood that two or more references to "an embodiment," "one embodiment," or "alternative embodiments" in various parts of this specification do not necessarily refer to the same embodiment. It should be noted that certain features, structures, or characteristics may be suitably combined in one or more embodiments of this specification.

[0183] It should be understood that in the above description of the embodiments of this specification, in order to facilitate understanding of a single feature, this specification combines various features into a single embodiment, drawing, or description thereof for the purpose of simplifying the specification. However, this combination of features is not essential, and those skilled in the art can easily extract some of the features and understand them as a single embodiment when reading this specification. In other words, the embodiments in this specification may be understood as a combination of multiple sub-embodiments. The content of each sub-embodiment may be valid even if it contains fewer than all of the features of the single embodiment disclosed above.

[0184] Each patent, patent application, patent application publication, and other material, e.g., article, book, specification, publication, document, article, etc., cited herein may be incorporated by reference. All content for all purposes, now or hereafter, is associated with this document, except for any claim history related thereto, any identical claim history that is inconsistent with or contradicts this document, or any identical claim history that has a limiting effect on the broadest scope of the claims. For example, in the event of any inconsistency or contradiction between the description, definition, and / or use of a term associated with any included material and the description, definition, and / or use of the related term in this document, the term in this document shall control.

[0185] Finally, it should be understood that the embodiments of the application disclosed herein are illustrative of the principles of the embodiments of the present specification. Other modified embodiments are also within the scope of the present specification. Therefore, the embodiments disclosed herein are merely examples and are not limiting. Those skilled in the art can implement the application of the present specification using alternative configurations based on the embodiments of the present specification. Therefore, the embodiments of the present specification are not limited to the embodiments exactly described in the application.

Claims

1. 1. A data processing method comprising: selecting an initial frame in the initial data that includes a predetermined number of bytes of initial data; performing data compression on the initial frame to obtain a compressed frame; wherein the data compression includes edge adjustment of a frame under compression, and the frame under compression includes the initial frame and any data state of the initial frame before it becomes the compressed frame in a data compression process; Here, the edge adjustment is performed by determining whether the amplitude of each unit among the multiple units of the frame being compressed in a target frequency region is large or small, and adjusting the amplitude in the target frequency region using a corresponding edge adjustment coefficient, thereby decreasing or increasing the amplitude, wherein the target frequency region is a mid- to high-frequency region, and the edge adjustment coefficient is greater than 0.

2. As mentioned above, edge adjustment is performed on frames during compression. Dividing the frame being compressed into the plurality of units based on a preset unit size; and adjusting the amplitude of each of the units in the target frequency region using the corresponding edge adjustment coefficients to increase or decrease the amplitude of the frame being compressed in the target frequency region.

2. The data processing method according to claim 1.

3. The above-mentioned adjustment to increase or decrease the amplitude of each unit in the target frequency region using the corresponding edge adjustment coefficient may include, for each unit: Selecting a function from a group of preset encoding functions as an encoding function, and adjusting them by the encoding function to obtain first units, so that components in the low frequency region within the frequency region are maintained but components in the target frequency region are attenuated; Calculating the difference between them and the first unit to obtain a first edge corresponding to them, wherein the first edge includes components in the target frequency region; adjusting the amplitudes of the first edges using the corresponding edge adjustment factors to obtain corresponding coded edges; 3. The data processing method according to claim 2, further comprising: convolving said first unit with said coded edge.

4. The aforementioned adjusting the amplitude of the first edge using the corresponding edge adjustment coefficients includes: determining that the edge value of the first edge is less than a predetermined first threshold, and increasing the amplitude of the first edge by the edge adjustment factor greater than 1; or determining that the edge value of the first edge is greater than a second predetermined threshold, and reducing the amplitude of the first edge by the edge adjustment factor less than 1; or 4. The data processing method according to claim 3, further comprising: obtaining the edge adjustment coefficient based on an optimization algorithm, with a weighted value of a distortion rate and a coding rate as an optimization target; and adjusting the amplitude of the first edge using the edge adjustment coefficient.

5. Increasing the amplitude of the first edge by the edge adjustment factor greater than 1 as described above may include: selecting one coefficient from a preset first group of edge adjustment coefficients as the edge adjustment coefficient to enhance the amplitude of the first edge, wherein all coefficients in the first group of edge adjustment coefficients are greater than 1; The aforementioned reducing the amplitude of the first edge by the edge adjustment factor less than 1 may include:

5. The data processing method according to claim 4, further comprising: selecting one coefficient from a preset second group of edge adjustment coefficients as the edge adjustment coefficient to reduce the amplitude of the first edge; and wherein all coefficients in the second group of edge adjustment coefficients are smaller than 1.

6. The aforementioned data compression of the initial frame is a method of first performing the edge adjustment on the initial frame, and then predicting the edge-adjusted initial frame to obtain a residual; a method of first performing a prediction on the initial frame to obtain a predicted frame, and then performing the edge adjustment on the initial frame and the predicted frame to obtain a residual; and first performing a prediction on the initial frame, obtaining a residual, and then performing the edge adjustment on the residual; 4. The data processing method of claim 3, wherein the compressed frame further includes the encoding function and the edge adjustment coefficient corresponding to each unit of the plurality of units.

7. 1. A data processing system comprising: at least one storage medium having stored thereon at least one instruction set for processing data; at least one processor communicatively coupled to the at least one storage medium; Here, when the data processing system is operating, the at least one processor reads the at least one instruction set and executes the data processing method described in any one of claims 1 to 6 in accordance with the instructions of the at least one instruction set.

8. 1. A data processing method comprising: obtaining compressed data, the compressed data including a compressed frame obtained by performing data compression on an initial frame, the data compression including edge adjustment; performing data decompression on the compressed frame to obtain a decompressed frame, the data decompression including edge compensation on the decompressed frame, the decompressed frame including the compressed frame and any data state of the compressed frame before it becomes the decompressed frame in the data decompression process; wherein the edge compensation and the edge adjustment have a predetermined correlation; The frame being compressed includes a plurality of units being compressed, and the edge adjustment includes adjusting the amplitude in the target frequency region of each unit being compressed among the plurality of units being compressed in the frame being compressed to decrease or increase the amplitude using a corresponding edge adjustment coefficient based on a determination of whether the amplitude in the target frequency region of each unit being compressed is large or small, wherein the target frequency region is a mid-frequency to high-frequency region, and the frame being compressed includes an original frame during data compression and data in an arbitrary state before the original frame becomes a compressed frame, and the edge adjustment coefficient is greater than 0. A data processing method comprising:

9. the adjusting to decrease or increase the amplitude of the plurality of units being compressed decreases or increases the amplitude of the frame being compressed in the target frequency region; 9. The data processing method of claim 8, wherein the edge compensation includes compensating the amplitude of each of a plurality of decompression units of the decompression frame in the target frequency region using an edge compensation coefficient corresponding to the edge adjustment coefficient based on the correlation.

10. As mentioned above, edge adjustment is performed on frames during compression. Dividing the frame being compressed into the plurality of units being compressed based on a preset unit size; and adjusting the amplitude of each of the compressed units in the target frequency region using the corresponding edge adjustment coefficients, wherein for each of the compressed units: Selecting a function from a group of preset encoding functions as an encoding function, and adjusting them by the encoding function to obtain first units, so that components in the low frequency region within the frequency region are maintained but components in the target frequency region are attenuated; Calculating the difference between them and the first unit to obtain a first edge corresponding to them, wherein the first edge includes components in the target frequency region; adjusting the amplitudes of the first edges using the corresponding edge adjustment factors to obtain corresponding coded edges; and overlapping the first unit and the coded edge; The edge compensation performed on the decompressed frame as described above is Dividing the decompression frame into the plurality of decompression units based on a preset unit size; and compensating the amplitude of each of the expanding units in the target frequency region using the edge compensation coefficient corresponding to the edge adjustment coefficient.

11. The above-described compensation of the amplitude of each of the extension units in the target frequency region using the edge compensation coefficient corresponding to the edge adjustment coefficient includes, for each of the extension units, determining a decoding function and adjusting them by said decoding function to obtain second units so that components in the low frequency range within their frequency range are preserved but components in the target frequency range are attenuated; Calculating the difference between them and the second unit to obtain second edges corresponding thereto, wherein the second edges include components in the target frequency region; compensating the amplitude of the second edges using the edge compensation coefficients corresponding to the edge adjustment coefficients to obtain corresponding decoded edges; 11. The data processing method of claim 10, further comprising overlapping the decompressed unit and the decoded edge.

12. As mentioned above, determining the decoding function is selecting a function from a preset set of decoding functions as the decoding function; or 12. The data processing method according to claim 11, further comprising selecting, as the decoding function, a function corresponding to the encoding function from the preset group of decoding functions.

13. The step of compensating the amplitude of the second edge using the edge compensation coefficient corresponding to the edge adjustment coefficient includes: Selecting one coefficient from a preset group of edge compensation coefficients as the edge compensation coefficient to compensate the amplitude of the second edge; or 12. The data processing method of claim 11, further comprising determining the edge compensation coefficient based on a correlation between the edge adjustment coefficient and the edge compensation coefficient to compensate for the amplitude of the second edge.

14. The correlation is The edge compensation causes the amplitude of the extended frame at any frequency in a low to mid frequency range to be 85% or more of that of the initial frame; 11. The data processing method of claim 10, further comprising: the edge compensation causing a gentle increase in amplitude in a mid-frequency region of the decompressed frame relative to the initial frame; or the edge compensation causing a gentle decrease in amplitude in a high-frequency region of the decompressed frame relative to the initial frame.

15. 1. A data processing system comprising: at least one storage medium having stored thereon at least one instruction set for processing data; at least one processor communicatively coupled to the at least one storage medium; wherein, when the data processing system operates, the at least one processor reads the at least one instruction set and executes the data processing method according to any one of claims 8 to 14 in accordance with the instructions of the at least one instruction set.

Images