Intelligent quantification system and method
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHANGHAI SHENGWANG TECH CO LTD
- Filing Date
- 2025-02-21
- Publication Date
- 2026-06-12
Smart Images
Figure CN122205084A_ABST
Abstract
Description
Cross-reference to related applications
[0001] This application claims priority to U.S. Patent Application No. 18 / 976,029, filed on December 10, 2024. Technical Field
[0002] This invention relates to video compression technology in general, and mainly to intelligent quantization methods, computer programs and systems. Background Technology
[0003] Video compression standards are primarily used to reduce bandwidth and video file size while maintaining high video quality. The existing High Efficiency Video Coding (HEVC) is a video compression standard that achieves higher data compression at the same or similar bitrate while maintaining comparable video quality compared to High-Level Video Coding (AVC). HEVC employs integer discrete cosine transform (DCT) for varying image patch sizes and discrete sine transform (DST) for 4x4 patch sizes. Essentially, this standard compares different parts of a video frame, identifies redundant regions within a single frame or between consecutive frames, and then replaces these redundant regions with brief descriptions instead of the original pixels.
[0004] A key feature of HEVC is its use of motion vector (MV) prediction. MV is a form of motion estimation used to describe the transformation from one 2D image to another. This typically occurs between adjacent frames in a video sequence. Motion vectors can be associated with the entire image (global motion estimation) or a specific portion, such as a rectangular block or arbitrary block, even at the pixel level. In HEVC, a motion vector refers to a 2D vector used for inter-frame prediction, providing the offset between the coordinates of the current image and the coordinates of the reference image.
[0005] In the current HEVC standard, to balance the precision and encoding cost of the video MV, a precision of one-quarter of the MV's pixel is used. More advanced video coding standards may use a higher precision MV.
[0006] After prediction is performed on a given pixel block, a residual will exist. This residual must be processed to obtain a high-quality image. Transformation and quantization are used to compress the residual signal. Traditional quantization uses a predetermined quantization step size or quantization matrix. Transform functions are used to convert an image / frame from the pixel domain to the frequency domain. For example, the Discrete Cosine Transform (DCT) is a technique that applies pixel-domain methods to an image, converting the pixel domain to the frequency domain and thus compressing redundancy. In video coding standards, DCT is applied to the residual of each pixel block.
[0007] Quantization is the inevitable result of representing numerical values as numbers with a fixed number of decimal places. Based on the DCT coefficients, the quantization scaling code is divided element-wise by the quantization matrix, and each resulting element is rounded. The quantization parameter determines the step size that associates the transform coefficients with a finite set of steps. This value is proportional to the compression ratio. Quantization and inverse quantization formulas are applied to the transformed signal. After performing quantization and inverse quantization, an inverse transform operation is performed on the DCT coefficient block. Rate-distortion costs cannot be optimized using a predetermined quantization step size or quantization matrix.
[0008] Minimizing rate distortion costs and ensuring optimal quantization are of great importance; therefore, this invention proposes a system and method for intelligent quantization. Summary of the Invention
[0009] This system and method relate to the field of video compression, primarily focusing on intelligent quantization techniques during video encoding. This system and method can effectively reduce rate-distortion costs in encoded video frames.
[0010] According to some embodiments, the present invention provides a method and system for intelligent quantization. In this system and method, the rate distortion cost of video encoding is first estimated. An offset, determined based on a frequency-dependent variable and a quantization step size, is applied to the reconstruction process. The frequency-dependent variable is then transmitted to the decoder to derive the offset. Furthermore, the transform coefficients can be estimated as a Laplacian or Cauchy distribution with a mean of zero and a frequency of α, incorporating content-dependent parameters. The offset at the encoder can be modified based on the frequency and content-dependent variables. Modifying the offset minimizes overall distortion by moving the quantization step size closer to the center point of the Laplacian or Cauchy distribution of each quantization chamber. Additionally, the size of the quantization chambers can be varied to optimize distortion, and the system can determine which frequency coefficients to use. If a frequency coefficient is equal to or below a threshold, it is quantized to zero and not used; if the frequency coefficient is above the threshold, the quantization chamber size and frequency-dependent variables are transmitted to the decoder. In some embodiments, a Laplacian or Cauchy distribution is used to estimate the distortion function, and the entropy of each quantization chamber is used to estimate the rate function.
[0011] It is worth noting that the various functions of the present invention described above can be implemented individually or in combination. The functions of the present invention will be described in more detail below with reference to the accompanying drawings. Attached Figure Description
[0012] To more clearly illustrate the present invention, some embodiments of the present invention will be described below with reference to the accompanying drawings, wherein:
[0013] Figure 1 This is an example block diagram of a system for encoding and transmitting video content, drawn according to an embodiment of the present invention;
[0014] Figure 2 This is an example block diagram illustrating the logical stages employed when encoding video according to an embodiment of the present invention;
[0015] Figure 3 This is an example diagram of quantization step sizes for transform coefficients with different offsets, drawn according to an embodiment of the present invention.
[0016] Figure 4 This is an example diagram of the Laplace and Cauchy distributions of the transform coefficients drawn according to an embodiment of the present invention;
[0017] Figure 5 This is an example process flowchart of video signal transformation and quantization drawn according to an embodiment of the present invention;
[0018] Figure 6 This is a flowchart of an example sub-process of intelligent quantization drawn according to an embodiment of the present invention;
[0019] Figure 7A and 7B This is a schematic diagram of a computer system capable of intelligent quantization, drawn according to an embodiment of the present invention. Detailed Implementation
[0020] The present invention will be described in detail below with reference to several embodiments shown in the accompanying drawings, wherein some technical details will be described to facilitate a comprehensive understanding of the embodiments of the present invention. However, those skilled in the art can also implement the embodiments of the present invention without these specific details. On the other hand, well-known technical steps and / or structures will not be described in detail to avoid unnecessarily obscuring the present invention. The functionality and advantages of the embodiments can be better understood with reference to the following drawings and descriptions.
[0021] The accompanying drawings and descriptions below will help to understand the various aspects, functions, and advantages of exemplary embodiments of the present invention. Those skilled in the art should understand that the embodiments of the present invention described herein are presented by way of example only and are not intended to limit. All functions disclosed herein can be replaced by other functions having the same or similar purpose unless explicitly stated otherwise. Therefore, other modified embodiments also fall within the scope of the invention and its equivalents as defined herein. Therefore, the imperative and / or sequential terms used in this article, such as “will,” “will not,” “should,” “should not,” “must,” “must not,” “first,” “initially,” “next,” “subsequently,” “before,” “after,” “finally,” and “end,” etc., are not intended to limit the scope of the invention, as the embodiments disclosed herein are merely illustrative examples.
[0022] This invention relates to a system and method for implementing intelligent quantization during video content encoding. As shown in the figure... Figure 1 This is an example diagram of a High Efficiency Video Coding (HEVC) system, represented by 100. The coding standard aims to achieve the highest possible coding efficiency. Coding efficiency refers to the ability to encode video at the lowest possible bitrate while simultaneously achieving a quality threshold. Encoder system 102 uses intra-frame prediction to segment the incoming image into block-like regions, which serve as the first image frame or the first frame of a random access point. Intra-frame prediction is a method of predicting blocks / pixels in a given frame using other pixels within the same frame. After predicting the first frame using intra-frame prediction, inter-frame prediction techniques can be used to predict other frames. Inter-frame prediction refers to predicting block content based on data from adjacent frames. After prediction, the image passes through a loop filter, and the final image is stored in a decoded image buffer. The image stored in the decoded image buffer can be used to predict other images.
[0023] In this system, multiple sub-components of the encoding and transmission module 102 receive video input 110. These sub-components include a general encoder 120 and transform, scalar, and quantizer 130, intra-frame predictor 143, and inter-frame predictor 155. The general encoder 120 generates general control data, which is transmitted to header formatting and context-adaptive binary arithmetic coding (CABAC) for processing and then merged into the encoded bitstream. The general control data is also transmitted to transform, scalar, and quantizer 130, intra-frame predictor 143, and inter-frame predictor 155 (not shown).
[0024] Transformer, scalar, and quantizer 130 performs scalar and transform operations on the input video frame and provides the output as quantized transform coefficients to the header formatting and context-adaptive binary arithmetic coding (CABAC) algorithm for incorporation into the encoded bitstream. The output data is also transmitted to scalar and inverse transformer 170. Transform units of different sizes can be applied to encode the prediction residuals. These transform units can use discrete cosine transform or discrete sine transform. Then, scalar and inverse transformer 170 provides the output to deblocker and filter module 180, transmits the output data to deblocking and filtering module 180, and to intra-predictor 143 and intra-predictor 145.
[0025] The transform, scalar, and quantizer 130 are components that perform intelligent quantization. During intelligent quantization, an offset is selected to minimize rate distortion cost. In some implementations, the rate distortion cost is derived from the following formula: min(D+λR) Formula 1: Rate-distortion cost function
[0026] In Equation 1, D represents the distortion, R represents the rate, and λ represents the Lagrange multiplier. The quantization process is performed according to the following formula: Formula 2: Quantization function
[0027] The formula for calculating the dequantization value is: W′=ZΔ Formula 3: Dequantization function
[0028] In formulas 2 and 3 above, f is the offset, W is the transform coefficient, Δ is the quantization step size, and W' is the dequantization coefficient.
[0029] Intra-frame estimator 143 uses various prediction algorithms to estimate the pixel values of adjacent pixels within the same frame. The output of intra-frame estimator 143 is transmitted to intra-frame predictor 145, which uses the estimated values to generate predictions for the relevant pixels. Conversely, inter-frame estimator 155 receives adjacent frame data from decoded image buffer 190 and estimates the motion trajectory between adjacent frames. The output of the motion estimation is transmitted to inter-frame compensator 153, along with header formatting and CABAC, and finally merged into the encoded bitstream (not shown).
[0030] Inter-frame compensator 153 generates motion compensation information. Selector 160 selects between intra-frame predicted image data and inter-frame motion compensation data. This information is fed back to transform, scalar, and quantizer 130, as well as deblocking and filtering module 180 (not shown).
[0031] The deblocking and filtering module 180 generates filtering control data, which is transmitted to header formatting and CABAC and finally merged into the encoded bitstream (not shown). The deblocking and filtered data is also transmitted to the decoded image buffer 190. The output of the decoded image buffer 190 includes output video 199.
[0032] refer to Figure 2 The diagram illustrates the logical flow and data conversion of generating the bitstream 290 from the original video 210. First, a subtraction operation is performed on the original video 210. Subtraction involves dividing the frame into blocks of one or more sizes. In some embodiments, the block size ranges from 4x4 to 64x64 pixels. Next, a two-dimensional discrete cosine transform (DCT) 220 is performed on each block. DCT can significantly reduce the memory and bandwidth required for compressed video. DCT 220 is applied to each residual value, including residual values from intra-frame coding and inter-frame coding.
[0033] Following DCT 220, the output is provided to quantization module 230. The quantization scaling code divides element-wise by the quantization matrix and rounds each calculation result. The quantization parameters determine the step size associated with a set of finite step sizes for the transformed coefficients. Next, the residuals are reconstructed using inverse quantization 240 and inverse DCT 250, respectively. The resulting residual blocks are then recombine with the motion compensation results from 270 using an addition function.
[0034] Motion estimation 260 uses the output of deblocking and the original video 210 to encode one frame for another. Motion estimation 260 encodes frame data by modifying the form of adjacent frames. The goal of motion estimation is to find the best match between two adjacent frames. The input to motion estimation is a macroblock and a search region. Motion estimation 260 performs block motion estimation, using a search algorithm to compute motion vectors (MVs). The most basic search method is to use a full search algorithm, which processes all pixels within the search range and finds the best block match through a cost function. The output of motion estimation is provided to motion compensator 270, which in turn uses it for addition operations. Furthermore, the output of motion estimation, along with the output of the quantization step, is transmitted to entropy encoder 280.
[0035] The entropy encoder 280 is a lossless data compression scheme. It creates and assigns a unique prefix code to each unique symbol in the input. The quantization result of each macroblock is entropy encoded to generate the bitstream 290.
[0036] refer to Figure 3 , Figure 3 This is an example of a quantization process. In this example diagram, there are two lines representing the reconstruction process W'. The upper line 310 represents the reconstruction process for intra-frame prediction, and the lower line represents the reconstruction process for inter-frame prediction. For these reconstruction processes, the representative value of W' is marked with a fixed interval Δ (quantization step size), as shown by the bubbles on lines 310 and 320. However, the parameter f has different values in inter-frame and intra-frame prediction. Therefore, the decision point (shown along the shorter line of the reconstruction process W') appears at different locations in intra-frame and inter-frame prediction. For example, assuming the parameter f in the reconstruction process W'310 above is Δ / 2, the decision point appears at half the value of W'. In contrast, in inter-frame prediction, as shown by line 320 below, the parameter f is Δ / 6, and the decision point appears closer to the representative value. The reconstruction process of related techniques is a fixed process, as shown in the following equation: W′=Δ×Z Formula 4: Fixed Reconstruction Process
[0037] Where Δ represents the interval of the values, and Z is the number of intervals. Currently, it is recommended to run the reconstruction process using offsets according to the following formula: W′=ZΔ+f' Formula 5: Offset Reconstruction Process
[0038] Where f' is determined by α and Δ. The parameter α depends on the frequency and content. This shifts the reconstructed value W' to the center point of the quantization bin (ZΔ-f, (Z+1)Δ-f). The offset f' can be derived by transmitting α to the decoder. In another embodiment, f' is a fixed value at each quantization frequency. The offset f' for each interval Z can be calculated using the following formula: Formula 6: Offset Calculation
[0039] Where p(x) is the probability distribution for each x, and x refers to each transformation coefficient. Let's look at... Figure 4 , Figure 4 These are example plots of two distributions: the Laplace distribution (represented by curve 420) and the Cauchy distribution (represented by curve 410). Transformation coefficients typically follow a distribution similar to the Laplace or Cauchy distribution. The Laplace distribution can be used to estimate the transformation coefficients. The Laplace distribution is derived from the following equation: Formula 7: Laplace Distribution
[0040] We can assume the mean of the distribution is zero. Optimal quantization should minimize the rate-distortion cost, as shown in Equation 1. Here, the Laplace distribution p(x) and the interval Δ can be used to estimate the distortion. The rate can be estimated using the entropy of each quantization chamber. For example, using this estimate, the distortion can be calculated using the following formula: Formula 8: Distortion Estimation
[0041] The rate can be estimated using the following formula: Formula 9: Rate Estimation
[0042] Where P i It is given by the following equation: Formula 10
[0043] The offset at encoder f can also be modified based on α for each frequency. By changing f, ZΔ can be brought closer to the center point, thereby minimizing overall distortion. This can be solved using the following formula: Formula 11: Offset Optimization Function
[0044] The quantization chamber size can also be changed to achieve optimal rate distortion. Given λ and α, the optimal quantization chamber size and the corresponding center of each quantization chamber can be determined. λ and α can be transmitted to the decoder to automatically derive inverse quantization. This is solved using the following equation: Formula 12: Optimal Quantitative Trading Container Size
[0045] Where B is the size of the quantitative trading platform, and: Formula 13: Quantitative ...
[0046] The rate R is derived from Equation 9, and the rate estimation is subject to the following conditions: Formula 14
[0047] Based on the method described above for optimizing the quantization bin size B, it is also possible to selectively determine which frequency coefficients to use. For high frequencies, most coefficients are quantized to zero, eliminating the need to transmit λ and α. Therefore, the system can send a signal to the decoder indicating which frequency coefficients will be quantized using the method described above.
[0048] Continue to refer to Figure 5 , Figure 5 This is an example flowchart of an intelligent quantization method, generally represented by 500. In this example method, prediction is first performed to generate residuals, as shown in 510. The residuals are then transformed using DCT or some alternative transform algorithm, as shown in 520. Finally, the residual signal is quantized at 530. Figure 6 A more detailed example of the quantization process is shown. First, distortion is estimated at 610 using Equation 8. Similarly, rate is estimated at 620 using Equation 9. At 630, the offset for the reconstruction process is selected by optimizing Equation 5. At 640, the offset is calculated using Equation 6. In another embodiment, the offset can be derived at the encoder using Equation 11. Next, the quantization chamber size is optimized at 650 using Equation 12. Finally, at 660, the system selects the frequency coefficients to be used with this quantization method. At 670, the residuals are quantized using these optimized parameters. This completes the entire process.
[0049] The above has described the systems and methods for intelligent quantization. Now let's look at the devices used to perform these functions in real time. For ease of discussion, Figure 7A and Figure 7B Each system is shown as computer system 700, which can be used to implement the embodiments of the present invention. Figure 7AThis diagram illustrates one possible physical form of the computer system 700. Of course, the computer system 700 can have various physical forms, ranging from printed circuit boards, integrated circuits, small handheld devices to large supercomputers. The computer system 700 may include a monitor 702, a display 704, a stand 706, a blade server containing one or more storage drives 708, a keyboard 710, and a mouse 712, etc. Medium 714 is a computer-readable medium used for transmitting data to the computer system 700. Figure 7B This is an example block diagram of computer system 700. System bus 720 connects to a variety of subsystems. Processor 722 (also called central processing unit or CPU) is adapted to storage devices (including memory 724). Memory 724 includes random access memory (RAM) and read-only memory (ROM). As is well known to those skilled in the art, ROM is used for unidirectional transfer of data and instructions to the CPU, while RAM is typically used for bidirectional transfer of data and instructions. Both types of memory can include any suitable form of computer-readable medium described below. Fixed medium 726 can also be bidirectionally adapted to processor 722 to provide additional data storage capacity and can also include any computer-readable medium described below. Fixed medium 726 is an auxiliary storage medium (e.g., hard disk) for storing programs, data, etc., and typically operates slower than main memory. It should be noted that, where appropriate, information stored in fixed medium 726 can be merged into memory 724 as virtual memory in a standard manner. Removable medium 714 can take the form of any computer-readable medium described below.
[0050] Processor 722 is also compatible with various input / output devices, such as display 704, keyboard 710, mouse 712, and speaker 730. Generally, input / output devices can be any of the following: video display, trackball, mouse, keyboard, microphone, touch-sensitive display, sensor card reader, tape or paper tape reader, tablet computer, stylus, voice or handwriting recognition device, biometric reader, motion sensor, EEG reader, or other computer, etc. Processor 722 can also interconnect with another computer or telecommunications network using network interface 740. It is conceivable that processor 722 uses network interface 740 to receive information from the network, or may output information to the network during the execution of the enhanced image patch prediction method described above. Furthermore, embodiments of the method of the present invention can run independently on processor 722, or can also operate collaboratively with a remote CPU sharing partial processing via a network such as the Internet.
[0051] Software is typically stored in non-volatile memory and / or drive units. In fact, for large programs, it may not even be possible to store the entire program in a single memory. However, it is understood that during software execution, it can be moved to a computer-readable location suitable for processing, which, for ease of explanation, is referred to as memory, if necessary. Even when the software is moved to memory for execution, the processor typically uses hardware registers and caches to store software-related values, ideally for speeding up execution. In this document, when a software program is stated to be “implemented in a computer-readable medium,” it is assumed that the software program is stored in any known or convenient location (non-volatile memory or hardware registers, etc.). A processor is considered “configured to run a program” when at least one value related to the program is stored in a processor-readable register.
[0052] During operation, the computer system 700 can be controlled by operating system software (such as a media operating system) that includes a file management system. For example, Microsoft Corporation in Redmond, Washington... An operating system series and its associated file management system is essentially operating system software with a file management system. For example, the Linux operating system and its associated file management system are also operating system software with a file management system. File management systems are typically stored in non-volatile memory and / or drive units, allowing the processor to perform various operations required by the operating system, input and output data, and store data in memory, including storing files in non-volatile memory and / or drive units.
[0053] Some parts described in detail in this document may be presented in the form of algorithms and symbolic representations of operations on data bits within computer memory. These algorithmic descriptions and representations are the means by which those skilled in the art of data processing most effectively communicate their work to others skilled in the art. As defined herein, an algorithm is a series of self-consistent operational steps designed to achieve a desired result. These operations require physical manipulation of physical quantities. Typically, these physical quantities are represented as electrical or magnetic signals that can be stored, transmitted, combined, compared, or otherwise manipulated, but this is not always necessary. For common use and ease of interpretation, these signals are often referred to as bits, values, elements, symbols, characters, items, numbers, etc.
[0054] The algorithms and representations described herein are inherently independent of any particular computer or device. The procedural methods described herein can be implemented using various general-purpose systems, or specific embodiments can be designed for dedicated devices to run, thereby achieving greater convenience. The architectures required for these systems will be detailed below. Furthermore, the techniques described herein are not referred to in any particular programming language, and therefore various embodiments can be implemented using various programming languages.
[0055] In other embodiments, the machine may operate as a standalone device or may be interconnected (e.g., networked) with other machines. In a networked deployment, the machine may operate as a server or client machine in a client-server network environment, or as a peer machine in a peer-to-peer (or distributed) network environment.
[0056] The aforementioned machines can be server computers, client computers, personal computers (PCs), tablets, laptops, set-top boxes (STBs), personal digital assistants (PDAs), cellular phones, iPhones, BlackBerry phones, processor-equipped glasses, processor-equipped headphones, virtual reality devices, processors, collaborative distributed processors, telephones, network devices, network routers, switches, or bridges, or any machine capable of executing a set of instructions (whether sequentially or otherwise) specifying the operations that the machine needs to perform.
[0057] Although machine-readable media or machine-readable storage media are shown as a single medium in exemplary embodiments, "machine-readable media" and "machine-readable storage media" should be understood to include a single medium or multiple media (e.g., a centralized or distributed database and / or associated caches and servers) that store one or more sets of instructions. "Machine-readable media" and "machine-readable storage media" should also be understood to include any medium capable of storing, encoding, or carrying a set of instructions for machine execution and enabling the machine to perform any one or more methods of the currently disclosed technical solutions.
[0058] Generally, routines that run to implement embodiments of the present invention may be implemented as part of an operating system or a particular application, component, program, object, module, or sequence of instructions (referred to as a "computer program"). A computer program typically includes one or more instructions written at different times in various memories and storage devices within a computer (or distributed across computers), and when one or more processing units or processors within (or across computers) read and execute these instructions, the computer can perform operations to implement the functions disclosed in the present invention.
[0059] Furthermore, while the embodiments described herein are set in the context of fully functional computers and computer systems, those skilled in the art will understand that various embodiments can be distributed in various forms of program products, and the disclosure of this invention applies equally to any particular type of machine or computer-readable medium in which the distribution is actually implemented.
[0060] While the invention has been described through several embodiments, variations, modifications, substitutions, and equivalents that fall within the scope of the invention are still possible. Although subsection headings have been used herein to aid in the description of the invention, these headings are illustrative only and are not intended to limit the scope of the invention. It should also be noted that many alternatives are available for implementing the methods and apparatus of the invention. Therefore, the appended claims should be considered to encompass all such variations, modifications, substitutions, and equivalents within the spirit and scope of the invention.
Claims
1. A computerized method for intelligent quantification, characterized in that, include: Estimating the rate-distortion function in video coding; An offset is applied to the reconstruction process, wherein the offset is determined based on a frequency-related variable and a quantization step size; as well as The offset is obtained by transmitting frequency-related variables to the decoder.
2. The computerized method for intelligent quantization according to claim 1, characterized in that, It also includes estimating the transformation coefficients as a Laplace distribution or a Cauchy distribution, wherein the Laplace distribution or Cauchy distribution has a mean of zero, a frequency of α, and includes at least one content-related parameter.
3. The computerized method for intelligent quantization according to claim 2, characterized in that, It also includes modifying the encoder offset based on frequency and at least one content-related variable.
4. The computerized method for intelligent quantization according to claim 3, characterized in that, Modifying the offset minimizes overall distortion by moving the reconstructed values in the quant container to the center of the Laplace or Cauchy distribution.
5. The computerized method for intelligent quantization according to claim 1, characterized in that, This also includes changing the size of the quantitative trading platform to optimize distortion.
6. The computerized method for intelligent quantization according to claim 5, characterized in that, It also includes determining which frequency coefficients to use.
7. The computerized method for intelligent quantization according to claim 6, characterized in that, The frequency coefficient is quantized to zero when it is equal to or below the threshold and will not be used; while when the frequency coefficient is above the threshold, the quantization chamber size and frequency-related variables are transmitted to the decoder.
8. The computerized method for intelligent quantization according to claim 2, characterized in that, The distortion function is estimated using a Laplace distribution or a Cauchy distribution, and the rate function is estimated using the entropy of each quantization chamber.
9. The computerized method for intelligent quantization according to claim 8, characterized in that, The quantification is derived from the following formula: W′=ZΔ Where W is the transform coefficient, Δ is the quantization step size, f is the offset, and Z is an integer.
10. The computerized method for intelligent quantization according to claim 9, characterized in that, The offset during the reconstruction process is: W' = ZΔ + f'.
11. A video encoding computer system capable of intelligent quantization, characterized in that, include: The image segmentation module is used to segment video images into blocks; The conversion module is used to convert individual image blocks; as well as The quantization module is used to estimate the rate-distortion function in video coding and apply the offset to the reconstruction process. The offset is determined based on frequency-related variables and the quantization step size, and is obtained by transmitting the frequency-related variables to the decoder.
12. The video encoding computer system according to claim 11, characterized in that, The quantization module is configured to estimate the transformation coefficients as a Laplace or Cauchy distribution with a mean of zero, a frequency of α, and containing at least one content-related parameter.
13. The video encoding computer system according to claim 12, characterized in that, The quantization module is also configured to modify the encoder offset based on frequency and at least one content-related variable.
14. The video encoding computer system according to claim 13, characterized in that, Modifying the offset minimizes overall distortion by moving the reconstructed values in the quant container to the center of the Laplace or Cauchy distribution.
15. The video encoding computer system according to claim 11, characterized in that, The quantitative module is also configured to change the size of the quantitative bin to optimize distortion.
16. The video encoding computer system according to claim 15, characterized in that, The quantization module is further configured to determine which frequency coefficients to use.
17. The video encoding computer system according to claim 16, characterized in that, The frequency coefficient is quantized to zero when it is equal to or below the threshold and will not be used; while when the frequency coefficient is above the threshold, the quantization chamber size and frequency-related variables are transmitted to the decoder.
18. The video encoding computer system according to claim 12, characterized in that, The distortion function is estimated using a Laplace distribution or a Cauchy distribution, and the rate function is estimated using the entropy of each quantization chamber.
19. The video encoding computer system according to claim 18, characterized in that, The quantification is derived from the following formula: W′=ZΔ Where W is the transform coefficient, Δ is the quantization step size, f is the offset, and Z is an integer.
20. The video encoding computer system according to claim 19, characterized in that, The offset during the reconstruction process is: W′=ZΔ+f′.