56results about How to "Improve compression ratio" patented technology

A method and an apparatus are provided for context-based adaptive binary arithmetic coding (CABAC) based encoding or decoding using improved selection of a context model for a high compression ratio. The method of CABAC-based encoding includes selecting a first context model based on statistics of previously input symbols and a second context model having a value of a most probable symbol (MPS) different from that of the first context model, encoding a current input symbol using the first context model if the current input symbol has a same value as that of the most probable symbol of the first context model or if the second context model cannot be selected, and encoding the current input symbol using the second context model if the current input symbol has a value different from the most probable symbol of the first context model and if the second context model can be selected.

The invention discloses an image-library-based video and image coding and decoding method, which comprises the following steps of: comparing multimedia data with pre-stored data in a coding database to obtain data, different from and identical with the pre-stored data, in the multimedia data; acquiring the attribute information of the identical data from the coding database; coding the different data and the attribute information of the identical data to obtain a data stream; and transmitting the data stream. In the method and a system, the multimedia data is compared with the pre-stored datato obtain the data, different from and identical with the pre-stored data, in the multimedia data, the different data and the attribute information of the identical data is coded, and the coding is not required by the whole multimedia data and only by the different data and the attribute information; and for the integral multimedia data, the different data forms a small part and the identical data forms the majority, and the coding of the identical data is performed for the attribute information of the identical data, so the volume of coded data is small, and a compression ratio is higher.

The invention discloses an image compression framework combining super-resolution and residual coding technology. The method includes establishing the mapping relationship of an original image and a sampling image compressed in different coding rate in an offline state, and constructing a super-resolution reconstruction model under each coding frequency; sampling the original image under iteration back projection to obtain the sampled small image; compressing the small image via a third-party encoder and transmitting the code stream and the de-noising auxiliary information; reconstructing the decoded small image into a big image at the decoding end by means of the super-resolution reconstruction model; suppressing the compressed noise in the big image through the combination with the de-noising auxiliary information; calculating the residual of the decoded big image and the original image at the coding end and transmitting the residual to the decoding end by means of the residual coding technology; and decoding the residual image at the decoding end, overlapping with the decoded big image, and recovering the final decoded image. An experiment shows that the framework provided can realize the compression ratio higher than JPEG2000.

A method for compressing 2-byte character data is provided to save a storage space by compressing a 2-byte text message and storing the compressed data. The maximum number of character strings, the number of code words and initial dictionary entry number are initialized, characters with high frequency are stored in a basic dictionary table, and a variable indicating a next codeword to be registered is initialized(S101). An additional compression-available codeword is stored in a supplementary dictionary table including the basic dictionary table, and the variable indicating the next codeword to be registered is re-initialized(S102). Message data is sequentially inputted and if the inputted data is included in the compression-available codeword, a corresponding code is searched from the dictionary table and outputted(S103). It is checked whether the corresponding code is in the dictionary(S104). It is determined whether it is the end of the data(S105). If so, a flushing process is performed(S106).

The invention relates to a compression implementation method for Beidou satellite information transmission. The method includes the steps of compiling a compression dictionary, and arranging content to be transmitted according to serial numbers or performing encoding through a binary system in the compression dictionary; splitting content after extraction, and when the content is Chinese, searching a Chinese dictionary to obtain corresponding Chinese words through matching, otherwise searching English, digital and symbol dictionaries; after word segmentation is performed on the content to be transmitted, inputting a final word segmentation result; performing indexing of the compressed dictionary on the split content, performing encoding to obtain a compressed short message and sending outthe compressed short message; and extracting received content, if it is judged that the content is not character table content, extracting one more byte backward, inquiring the compressed dictionary after forming two bytes, and after all byte content is converted, outputting and displaying all the byte content. According to the compression implementation method for Beidou satellite information transmission, through customization of the compressed dictionary, encoding compression is performed on the short message content, and expansion is performed within a limited byte number, thereby improving transmission speed of Beidou RD communication and increasing transmission content.

The invention relates to a compression method and device and a decompression method and device for an image with transparency. A compression method for the image with transparency comprises the steps of obtaining the image data of original channels A, R, G and B of the image with transparency; respectively pre-multiplying the image data of the channels R, G and B by the image data of the channel A, wherein a pre-multiplying formula is RGB=RGB*A/MAX (A), MAX (A) denotes the maximum value of the data of the channel A; saving the image data of the channels R, G and B obtained through pre-multiplication as a first nontransparent image and saving the image data of the channel A as a second nontransparent image, wherein a formula for the image data of the channel A is R=G=B=A; respectively compressing the first nontransparent image and the second nontransparent image by adopting a nontransparent image lossy compression algorithm. The compression method and device for the image with transparency provided by the invention have the advantage that the compression ratio of the image with transparency is improved.

The invention provides a positioning data compression method used for a Beidou vehicle-mounted system. The method includes the steps that N pieces of positioning data are sequentially converted into N pieces of preprocessed positioning data including the positioning state, the latitude, the longitude, the ground velocity and the ground azimuth, and the value of N is recorded with a preset block counter; half-byte compression is performed on the N pieces of preprocessed positioning data including the positioning state, the latitude, the longitude, the ground velocity and the ground azimuth, and N pieces of first compressed positioning data are obtained; the statement identifier header, the world standard time, the latitude hemisphere, the longitude hemisphere and the date of the first piece of stored positioning data and the value of N recorded by the preset block counter are added to the front end of the first piece of compressed positioning data among the N pieces of first compressed positioning data, and then the first piece of compressed positioning data among the N pieces of first compressed positioning data and the other N-1 pieces of first compressed positioning data form second compressed positioning data together. Through the positioning data compression method and system, the data compression ratio is increased, and system resource consumption is reduced.

The invention discloses a neural network pruning method based on combination of sparse learning and a genetic algorithm, and mainly solves the problem that a neural network consumes a large amount ofstorage resources and computing resources. The implementation scheme is as follows; training a neural network in a sparse learning mode to obtain a sparse network structure; carrying out heuristic search on potential sub-networks in the trained neural network by utilizing a genetic algorithm and a dynamically adjustable evaluation factor, and automatically searching an optimal sub-network meetingrequirements under the guidance of a fitness function; and retraining the obtained optimal sub-network to obtain a final pruning result. According to the method, the consumption of storage resources and computing resources by the neural network is reduced, the precision of the pruned network is ensured, adjustable pruning can be realized, and the method can be used for compressing the neural network.

The invention provides an FPGA-based image compression method and apparatus, and an FPGA-based image transmission system, and relates to the technical field of image processing. The FPGA-based image compression method comprises the following steps: receiving image data, and storing the image data in an idle subarea, and meanwhile marking the idle subarea as a cache subarea; when the image data in the cache subarea reach a set threshold, marking the cache subarea as an encoding subarea; performing encoding compression on the image data in the encoding subarea to generate image compression data, and marking the encoding subarea as a to-be-sent subarea; and sending the image compression data in the to-be-sent subarea according to a sending instruction. The existing technical problem that real-time lossless transmission of the image data cannot be realized is solved, and a technical effect of real-time lossless transmission of the image data is realized.

The invention relates to a picture compression method in an AOI (Automatic Optic Inspection) detection system. The picture compression method comprises the following steps: step 1, arranging a display panel defect detection image group, determining a pure white detection image in the display panel defect detection image group as a first detection image, and sorting other detection images except the pure white detection image in the display panel defect detection image group in any order; step 2, performing a difference value subtraction operation on pixel values of other images except the pure white detection image and a pure black detection image in the display panel defect detection image group and a pixel value of the first detection image orderly, and obtaining a plurality of corresponding foreground images; step 3, removing a negative value of the pixel value of each foreground image and obtaining each corrected foreground image; and step 4, respectively performing image lossless compression on each corrected foreground image, and directly performing image lossless compression on the pure black detection image and the pure white detection image. The picture compression method in the AOI detection system provided by the invention can guarantee the compression ratio and shorten the compression time.

The invention discloses a hierarchical tree index-based correlation data compression method. The method comprises the following steps of constructing a dictionary file; establishing a two-dimensionalmatrix tree index part for indexing data blocks; and establishing three-dimensional matrix indexes of block data and corresponding ID triple serialization data. According to the hierarchical tree index-based correlation data compression method, a relational three-dimensional matrix is subjected to block segmentation in combination with the relational three-dimensional matrix and predicate vector distribution; through upward mapping of the relational three-dimensional matrix, the indexes in a two-dimensional matrix form are established for the block data; and compression storage is performed oneach block, so that the compression rate is remarkably increased on the basis of not destroying an original structure.

The invention relates to an associated data compressing method friendly to query. The method comprises the following steps: defining a relation mining rule, and mining a potential incidence relation in a triad; defining a compression query memory model which consists of a subject vector, a predicate vector and an object matrix; defining a serialization mode of the compression query memory model, and implementing serialization and deserialization by using three auxiliary symbols; defining a query mode of executing SPARQL on the compression query memory model, querying a subject and a predicate by using a binary search method, and querying an object by using a linear traverse method; and defining a scheme for solving slow query caused by the over-large object matrix, and dividing a large data block into a plurality of small data blocks. Compared with most of existing compression schemes, an associated data set processed by the method has the characteristics that the compression ratio is increased, and SPARQL query operation can be carried out directly under the compression state.

The invention discloses a waterproof well wall compressible device and a construction process thereof. The contractible device comprises an annular U-shaped steel structure, a circular ring supporting plate and compressible fillers. The annular U-shaped steel structure is constituted by head-tail connection of a plurality of U-shaped steel structure arc sections. The circular ring supporting plate is further formed by head-tail connection of a plurality of arc-shaped supporting plate bodies and fixedly connected to the inner side of the opening end of the annular U-shaped steel structure. A cavity formed between the annular U-shaped steel structure and the circular ring supporting plate is filled with the compressible fillers. The structure is simple; construction is convenient; airtightness is good; the compression ratio is high, working performance is reliable, the engineering cost and engineering construction difficulty can be effectively reduced; and quite wide practicability in the technical field is achieved.

The invention provides methods of compressing and decompressing gene sequences and a device of compressing the gene sequence. The method of compressing the gene sequence includes: generating a variation reference sequence according to high-frequency variation information and a standard reference sequence; and compressing the to-be-processed gene sequence according to a matching result of the to-be-processed gene sequence and the variation reference sequence to obtain the compressed gene sequence. According to the above-mentioned methods of compressing and decompressing the gene sequences and the above-mentioned device of compressing the gene sequence, a rate of compression for the gene sequence can be increased, thus storage space of the gene sequence is reduced, and copying and transmission for the gene sequence are facilitated.

The invention discloses a method for compressing and densifying fast-growing cunninghamia lanceolata, which relates to the technical field of wood processing treatment and comprises the following steps of: (1) processing cunninghamia lanceolata wood; (2) preparing a cunninghamia lanceolata compression and densification pretreatment pre-solution; (3) carrying out compression and densification pretreatment on cunninghamia lanceolata; (4) balancing the moisture content of the wood before pressing; (5) compressing and densifying the cunninghamia lanceolata wood; and (6) carrying out later-stage thermal fixation. The cunninghamia lanceolata compression and densification method provided by the invention is simple in process and equipment, short in required time and relatively low in production cost. The density, the hardness and the elasticity modulus of the obtained compressed and densified cedarwood are increased compared with those of bare cunninghamia lanceolata and compressed and densified cunninghamia lanceolata which is not pretreated, and the moisture absorption recovery rate and the water absorption recovery rate are small.

The invention provides a method for compressing time sequence data. The method comprises the steps of writing obtained time sequence data into a temporary table; determining to-be-compressed time sequence data from the temporary table according to a query time range; compressing the to-be-compressed time sequence data, then writing the compressed time sequence data into a final table, judging whether a current query time range can be updated or not, if the current query time range can be updated, turning to the step of determining the to-be-compressed time sequence data from the temporary table after the query time range is updated; and if the current query time range cannot be updated, compressing a compression result of the time sequence data written into the final table. The invention also provides a method for querying the time sequence data. The method comprises the steps of obtaining a generation time range of to-be-queried time sequence data; comparing marking time in a query time range employed when first compression is carried out with the generation time range, and determining a storage position of the to-be-queried time sequence data according to a comparison result, wherein the storage position comprises the temporary table and/or the final table; and obtaining the to-be-queried time sequence data according to the determined storage position.

The embodiment of the application provides a method and device for encoding or decoding a three-dimensional data point set, which can simplify an encoding method, reduce the time overhead of encodingand decoding, and improve the compression rate. The method comprises the following steps: carrying out position coordinate coding on the position coordinate of at least one three-dimensional data point in the three-dimensional data point set to obtain a first binary code stream; according to the position coordinate sequence of the at least one three-dimensional data point after the position coordinate coding, carrying out binarization coding on the attribute of the at least one three-dimensional data point to obtain a second binary code stream; and carrying out entropy coding on the first binary code stream and the second binary code stream.

There is provided a method of encoding data (D1) in an encoder (20) to generate corresponding encoded data (E2); the method includes: (a) receiving the data (D1) to be encoded and analyzing sub-portions of the data (D1) to be encoded to determine one or more encoding algorithms which are to be applied to encode the sub-portions, wherein the one or more encoding algorithms include at least one interpolation algorithm; (b) computing one or more interpolation parameters for the at least one interpolation algorithm which are representative of data values of the sub- portion of the data (D1) to be encoded by the at least one interpolation algorithm; (c) encoding a remainder of the sub-portions of the data (D1) to be encoded using the one or more encoding algorithms; and (d) combining data generated in (b) and (c) to generate the encoded data (E2). Moreover, a corresponding method of decoding encoded data (E2) in a decoder (30) to generate corresponding decoded data (D3) includes: (i) analyzing the encoded data (E2) to determine one or more encoding parameters included in the encoded data (E2), and the one or more coding algorithms to which they relate, wherein at least one of the coding algorithms include at least one interpolation algorithm; (ii) computing data values of sub-portions of the encoded data (E2) by applying inverse algorithms to the one or more parameters and their associated data present in the encoded data (E2); and (iii) combining the computed data values of the sub-portions to generate the decoded data (D3). The methods are susceptible to being employed in an encoder, in a decoder and in codec. Moreover, the methods are beneficially implemented on computing hardware by executing one or more software products thereupon, wherein the software products are arranged to implement the methods.

The invention discloses an image/video coding method, device and system and a computer readable storage medium, and the method comprises the steps of obtaining a to-be-coded block in a current coding frame, and obtaining adjacent coded pixels of the to-be-coded block; setting a search template based on the to-be-coded block and/or the adjacent coded pixels; constructing a motion information candidate list of the to-be-coded block, wherein the motion information candidate list comprises a plurality of pieces of motion information; determining a search starting point by using a co-location block of the to-be-coded block in a reference frame or motion information of the motion information candidate list, and searching an optimal search block matched with the search template in a preset search range; obtaining an optimal matching block according to the optimal search block; and coding the to-be-coded block by using the optimal matching block to obtain a code stream of the to-be-coded block. According to the method, the matching block of the to-be-coded block is searched in the reference frame by using the search template, so that the compression rate of inter-frame coding can be further improved.

The invention discloses a video coding and compressing method, relating to the technical field of video coding. The video coding and compressing method comprises the following steps of: S1. pre-coding source video streaming, reestablishing all groups of pictures (GOP), and obtaining X GOPs and X I frames; S2. calculating the macro-block coincidence degrees of all the I frames, merging the I-frames with the macro-block coincidence degrees reaching a threshold value into a group, and obtaining N groups of I frames, wherein N is less than or equal to X; S3. calculating an optimizing and coding factor K of the source video streaming subjected to the S2, calculating a coding rate Z for coding the source video streaming according to the K; S4. coding the source video streaming according to the coding rate Z, taking sample frames in the predetermined N groups of I-frames as the I frames during coding, and taking the I frames serving as non-sample frames as P frames taking the sample frames in the same group as reference frames. According to the video coding and compressing method, the compression ratio of videos is increased.

The invention relates to a method for processing an image, the method comprising: providing (101) a set of key points from the image; describing (103) location information of the set of key points in form of a binary matrix; and scanning (105) the binary matrix according to a predetermined order, thereby creating a new representation of the location information of the set of key points.

For data to be compressed and having periodicity, two windows of identical size are set for two intervals in accordance with peaks appearing almost periodically and sample data is alternately rearranged between the windows of the identical size, thereby replacing the frequency of data having periodicity with approximately a half frequency without deteriorating reproducibility into original data, so that the replaced data of a low frequency is subjected to compression. This rearrangement processing may be applied to a compression having a characteristic that the compression ratio cannot be increased in a high frequency region, so as to increase the compression ratio without deteriorating the quality of the reproduced data obtained by decompression.

The invention provides a DFA (Deterministic Finite Automaton) compression method and apparatus, and a regular expression matching method and system. The DFA compression method comprises the steps of constructing a DFA according to a regular expression set, wherein the DFA comprises characters, states and transitions, corresponding to the characters, of the states, and corresponding to each character, the transition points to a state corresponding to the character from a current state; marking the states and the transitions according to the same transitions in the DFA, wherein the same transitions are transitions pointing to the same state; and compressing the DFA according to the marking of the states and the transitions. By marking the states and the state transitions in the DFA and compressing the DFA according to the marking of the states and the transitions, a relatively high compression rate can be achieved without influencing a search rate, thereby enabling the DFA to occupy less memory.

The invention provides a network packet header compression method and a network packet header compression device used for network performance monitoring. The method comprises: the position informationof a transmitting device of a current network data package is obtained; if it is judged that the location information indicates that the transmitting device is located at an access location in the network, the packet header of the current network data packet is compressed based on the obtained terminal server physical address list. The invention provides a network data packet header compression method and a network data packet header compression device for network performance monitoring, which compress based on the physical address or the IP address granularity of the network data packet header according to the position of collecting the network data packet, thereby realizing higher compression ratio. The invention also reduces the bandwidth and storage space consumption in the process ofcollecting, transmitting and storing network packet headers when network performance is monitored, and reduces the fault tolerance of decompression.

The invention discloses a polarization SAR image compression method based on multi-direction dictionary learning. The method mainly solves the problem that the quality of an image compressed and reconstructed through the existing technology is low. The method includes the implementation steps of firstly, inputting a set of polarization SAR four-channel images, and conducting asymmetrical three-dimensional wavelet transform; secondly, conducting sparse representation on coefficient matrixes, in different directions, of high-frequency sub-bands under different scales of all the channels after the discrete wavelet transform is conducted so as to obtain all sparse matrixes; thirdly, conducting quantizing and coding on the coefficient matrixes of all low-frequency sub-bands after the discrete wavelet transform is conducted to obtain low-frequency code streams; fourthly, conducting unified quantizing and coding on the sparse matrixes in different scales and different directions so as to obtain high-frequency code streams; fifthly, forming final code streams through the low-frequency code streams and the high-frequency code streams. By means of the method, redundancy between the channels can be effectively eliminated, the marginal information and contour information of images are better preserved, the quality of a compressed and reconstructed image is improved, and the method can be used for transmitting and storing polarization SAR images.

The invention provides a video encoding method and device based on a long-term reference frame. The method comprises a step of obtaining a short term reference macroblock data and long term reference macroblock data corresponding to a current encoding macroblock, a step of using a first interframe prediction mode to obtain the interframe prediction information of a short term reference macroblock, and using a second interframe prediction mode to obtain the interframe prediction information of a long term reference macroblock, a step of carrying out fusion processing on all encoding sub macroblocks of the current encoding marcoblock according to the interframe prediction information of the short term reference macroblock and the interframe prediction information of the long term reference macroblock, and obtaining the interframe prediction information of the current encoding macroblock, and a step of carrying out motion compensation processing on all encoding sub macroblocks of the current encoding macroblock and obtaining the interframe prediction encoding data of the current encoding macroblock. The device provides an execution module for the method. By applying the video encoding method and device, the video compression rate and encoding quality can be improved effectively.

The present invention acquires much accurate image information having no noise, without increasing load in the post-processing of the image information, a receiving apparatus 3 includes: a synchronization signal detector 34 that detects a horizontal synchronization signal and a vertical synchronization signal; an image processor 35 that performs an image generation process of each frame based on the horizontal synchronization signal and the vertical synchronization signal detected by the synchronization signal detector 34; and an image deletion controller 36a that controls to delete an image of the current frame, when a horizontal synchronization signal within one frame detected by the synchronization signal detector 34 is not continuously detected by a first predetermined number or more, or when a horizontal synchronization signal within one frame detected by the synchronization signal detector 34 is not detected by a second predetermined number or more, or when a vertical synchronization signal detected by the synchronization signal detector 34 is not continuously detected by a third predetermined number or more.

The embodiment of the invention discloses an image processing method, an electronic device and a computer readable storage medium, and relates to the technical field of data processing. The image processing method comprises the following steps: analyzing each traffic element in each frame of image one by one, judging whether the traffic elements are repeated in a plurality of frames of images which are adjacent in time or not, and marking the repeated traffic elements; and deleting redundant images in all the images, wherein all the traffic elements in the redundant images are marked as the repeated traffic elements. The method, the electronic device and the computer readable storage medium can be used for improving the universality of redundant image screening and the compression rate ofrepeated traffic elements in the image.

The embodiment of the specification provides an image processing method for realizing privacy protection. A private image is subjected to desensitization and compression processing at a terminal; andimage recognition is carried out at a server side according to the desensitized and compressed data. Frequency domain features of the private image are extracted through the terminal; preprocessing and screening are carried out on the frequency domain feature map, and the feature map more beneficial to image recognition is obtained in a self-adaptive mode. Compared with pure image compression, the provided method has the advantages: the compression rate in the image transmission process is improved, data privacy is effectively protected, the data processing amount is reduced, so that the effectiveness of image processing is improved.

The invention provides a preprocessing method and system for achieving similar image compression based on an FPGA, and belongs to the technical field of image processing. Calculating the average valueor median of the data of the pixel points at the same position in the plurality of similar images to form a model image; comparing each similar image with the model image, and calculating an absolutevalue of a difference value of data of pixel points at the same position; if the absolute value of the difference value corresponding to more than two continuous pixel points is less than or equal toa preset model threshold, outputting the position information of the first pixel point and the quantity information of the continuous pixel points; and if not, outputting the original image data. Anexternal FPGA chip is used for processing, the compression rate is high, the compression time is short, a computer can be released for other processing, compressed data can be further compressed through a traditional compression method, and the image compression rate is increased.