CN118413675B - Context-based progressive three-plane coding image compression algorithm and terminal equipment - Google Patents
Context-based progressive three-plane coding image compression algorithm and terminal equipment Download PDFInfo
- Publication number
- CN118413675B CN118413675B CN202410879262.3A CN202410879262A CN118413675B CN 118413675 B CN118413675 B CN 118413675B CN 202410879262 A CN202410879262 A CN 202410879262A CN 118413675 B CN118413675 B CN 118413675B
- Authority
- CN
- China
- Prior art keywords
- tensor
- context
- image
- super
- potential
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000007906 compression Methods 0.000 title claims abstract description 47
- 230000006835 compression Effects 0.000 title claims abstract description 43
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 31
- 230000000750 progressive effect Effects 0.000 title claims abstract description 16
- 230000009467 reduction Effects 0.000 claims abstract description 33
- 208000037170 Delayed Emergence from Anesthesia Diseases 0.000 claims abstract description 10
- 238000004364 calculation method Methods 0.000 claims abstract description 10
- 230000000007 visual effect Effects 0.000 claims abstract description 9
- 230000004913 activation Effects 0.000 claims abstract description 4
- 238000005516 engineering process Methods 0.000 claims description 6
- 239000000284 extract Substances 0.000 claims description 5
- 238000007670 refining Methods 0.000 claims description 4
- 238000013527 convolutional neural network Methods 0.000 claims description 3
- 238000011156 evaluation Methods 0.000 claims description 3
- 230000004927 fusion Effects 0.000 claims description 3
- 238000013139 quantization Methods 0.000 claims description 3
- 238000000034 method Methods 0.000 description 10
- 238000012545 processing Methods 0.000 description 5
- 230000005540 biological transmission Effects 0.000 description 4
- 238000004590 computer program Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000007547 defect Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013144 data compression Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/13—Adaptive entropy coding, e.g. adaptive variable length coding [AVLC] or context adaptive binary arithmetic coding [CABAC]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/20—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using video object coding
- H04N19/21—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using video object coding with binary alpha-plane coding for video objects, e.g. context-based arithmetic encoding [CAE]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/30—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using hierarchical techniques, e.g. scalability
- H04N19/36—Scalability techniques involving formatting the layers as a function of picture distortion after decoding, e.g. signal-to-noise [SNR] scalability
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/85—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using pre-processing or post-processing specially adapted for video compression
- H04N19/88—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using pre-processing or post-processing specially adapted for video compression involving rearrangement of data among different coding units, e.g. shuffling, interleaving, scrambling or permutation of pixel data or permutation of transform coefficient data among different blocks
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention discloses a context-based progressive three-plane coding image compression algorithm and terminal equipment, which specifically comprise the following steps: s1, compressing a video stream acquired by a visual camera through an MPEG algorithm, and acquiring a visual image with a coding format of H265; s2, sequentially converting the visual image X into a potential tensor Y and a super potential tensor Z through an encoder and a super encoder, and obtaining an average value and a standard deviation of the representation Y by using a super decoder; s3, evaluating values by a probability calculation module by using the average value, the standard deviation and the coded three planes; s4, designing a context-based rate reduction module, predicting the value of each tri-plane, fusing the values with a convolution layer through a residual block, and finally using an activation function; s5, designing a context-based distortion reduction module, and performing image reconstruction X after entropy decoding; and S6, finally, reconstructing the improved potential tensor image through a super decoder.
Description
Technical Field
The invention belongs to the technical field of image processing, and particularly relates to a context-based progressive three-plane coding image compression algorithm and terminal equipment.
Background
Image compression algorithms are a series of techniques for reducing the size of digital image files in order to reduce the storage space and transmission bandwidth requirements without losing image quality as much as possible. It includes two main methods, lossless compression and lossy compression. Lossless compression algorithms allow complete recovery of the original image data, suitable for scenes with very high requirements on image quality. Lossy compression, in turn, achieves higher compression rates by discarding some of the less visually sensitive information, and is widely used in applications requiring the storage and transmission of large amounts of image data. With the development of deep learning and artificial intelligence technology, the field of image compression is coming to be an emerging technology of context-based coding, generation of countermeasure networks and end-to-end learning, and the like, and these advances provide new possibilities for improving compression efficiency and image quality.
The core problem of image compression algorithms is to balance the contradiction between compression efficiency and image quality, and the challenge of keeping the original information as much as possible while reducing the amount of data. Efficient compression is often accompanied by some degree of data loss, which is particularly evident in lossy compression. Meanwhile, complex algorithm designs in the compression process may cause increased encoding and decoding time consumption, which poses challenges for real-time processing and resource-constrained devices. In response to these problems, a more compact progressive encoding scheme is proposed, aiming at gradually refining the compression process through intelligent encoding steps, thereby realizing higher compression ratio while maintaining image quality, and opening up a new path for the progress of image compression technology.
Disclosure of Invention
The invention aims to provide a context-based progressive three-plane coding image compression algorithm and terminal equipment, which can predict third-order probability more accurately by combining context information, thereby realizing more compact coding. Not only improves the compression efficiency, but also helps to preserve richer image details. In addition, the algorithm also comprises a distortion reduction module which can intelligently extract key tensor information from three planes, so that the image quality loss in the compression process is further reduced. This allows for a higher level of compression performance while maintaining high quality image reconstruction, overcoming the problem of incompatibility of conventional tri-planar coding with autoregressive context models.
In order to achieve the above object, the present invention provides a context-based progressive three-plane coded image compression algorithm, comprising the steps of:
s1, compressing an original video stream by adopting an MPEG algorithm, and obtaining a visual image coded into an H265 format;
s2, sequentially converting the visual image X into a potential tensor Y and a super potential tensor Z through an encoder and a super encoder, and using quantization The super decoder obtains the average value and standard deviation of the representation Y;
S3, evaluating the value by using the average value, the standard deviation and the coded three planes through a probability calculation module ;
S4, designing a context-based rate reduction module (Rate Reduction Model, RRM), predicting the value of each tri-plane and fusing the values with a convolution layer through a residual block, and finally determining by using an activation functionA value;
S5, designing a context-based distortion reduction module (Distortion Reduction Model, DRM) and using the bias hidden tensor to perform random after entropy decoding Reconstructing an image by using values in a range;
and S6, finally, reconstructing the improved potential tensor image through a super decoder.
As a further scheme of the invention: the step S1 comprises the following steps:
By adopting an MPEG video stream image compression algorithm, good video quality under high compression ratio and low bit rate is realized through motion compensation, discrete Cosine Transform (DCT) and Huffman coding technology, and particularly, the method is good at processing motion scenes and reducing time redundancy; in addition, the method overcomes the defects of other series of algorithms in the aspects of dynamic video compression efficiency and bandwidth adaptability, and provides a more flexible and efficient video transmission solution.
As a further scheme of the invention: the step S2 comprises the following steps:
the encoder extracts the key features in the image, and finally outputs a potential tensor Y by using the progressive extraction of the image features layer by layer of the convolutional neural network; the super encoder further refines and abstracts the features, further compresses information and the like to obtain a super potential tensor Z;
Furthermore, by Calculation of quantized super potential tensorThe average value M and standard deviation are then obtained using a super decoder, as shown in equation (1),
(1);
As a further scheme of the invention: the step S3 comprises the following steps:
Using the mean, standard deviation and encoded tri-plane by the probability calculation module WhereinC represents the number of channels, H and W represent the height and width, respectively;
Using entropy parameter M and standard deviation And the coded trihedral value is used for evaluationWherein entropy parameters are typically used to adjust and control the entropy of information during encoding, thereby affecting the loss of information in compression rate.
As a further scheme of the invention: the step S4 comprises the following steps:
by designing a context-based rate reduction module (Rate Reduction Model, RRM), predicting potential elements in raster scan order is eliminated, but each plane is predicted ,In addition, RRM module pairRefining the probability estimates in (2) to produce updated tensors,Bit ratio required in entropy codingLess, thereby improving the performance of the "rate-distortion" RD.
As shown in fig. 2, RRM is input from the inputs respectivelyAnd extracting features in context and fusing them by a residual block convolution layer, wherein the fusion tensor has a value equal to that of the residual block convolution layerThe same spatial resolution, but the channels areFour times that of an additively added item by channelAnd a scale item;
First, S is converted to B as shown in equation 2:
(2);
specifically, each element thereof is in In (C) and then willAdded toAnd sums it to produce an updated probability tensor by B modulation。
As a further scheme of the invention: the step S5 comprises the following steps:
The context-based distortion reduction module (Distortion Reduction Model, DRM) is designed to be used after entropy decoding, unlike non-progressive encoders, where the algorithm proposed herein can use a bias hidden tensor For any one ofReconstructing an imageThus, in decodingThe DRM module then uses the context to reduce errors, thereby reducing image distortion.
As shown in FIG. 3, the architecture of DRM is shown, in M andFor context, the DRM module shifts the tensorIs to be ofAs shown in formula (3):
(3);
Unlike RRM, DRM does not use AndAs a context, as it includesDue to the probability information of (2)Has been decoded and used for reconstruction,AndHardly provideIn addition, the DRM module is a regressor for reducing distortion, and the RRM is a classifier for reducing bit rate.
The DRM module is trained to minimize the loss as shown in equation (4):
(4);
As a further scheme of the invention: the step S6 comprises the following steps:
finally, the improved latent tensor is passed through the super decoder Image reconstruction is carried out to obtain。
To achieve the above object, another aspect of the present invention provides a terminal device, including a processor, a memory, and a program or instructions stored on the memory and executable on the processor; the processor implements the image compression algorithm described above when executing the program or instructions.
Compared with the prior art, the invention has the following beneficial effects:
context information integration: by developing a context-based probability reduction module, the algorithm can more accurately predict and estimate the cubic probability of potential elements; the integration of the context information provides a more accurate coding basis for the data of three planes, so that the coding process is more efficient and compact.
Intelligent tensor extraction: by designing a context-based distortion reduction module, the module can intelligently identify and extract critical potential tensor information from three planes; the process not only optimizes the data compression efficiency, but also remarkably improves the quality of the reconstructed image, and ensures the maximum reservation of image details in the compression process.
The improvement of the algorithm brings new performance improvement to the field of depth progressive image compression, and is particularly suitable for application scenes with high requirements on compression efficiency and image quality.
Drawings
Fig. 1 shows a flowchart of an overall image compression method according to an embodiment of the present invention;
FIG. 2 illustrates a schematic diagram of a context-based rate reduction module provided by an embodiment of the present invention;
FIG. 3 is a schematic diagram of a context-based distortion reduction module according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of a terminal device according to an embodiment of the present invention.
Detailed Description
The invention is further illustrated by the following examples.
As shown in fig. 1, a context-based progressive tri-planar coded image compression algorithm includes the steps of:
s1, compressing an original video stream by adopting an MPEG algorithm, and obtaining a visual image coded into an H265 format;
s2, sequentially converting the visual image X into a potential tensor Y and a super potential tensor Z through an encoder and a super encoder, and using quantization The super decoder obtains the average value and standard deviation of the representation Y;
S3, evaluating the value by using the average value, the standard deviation and the coded three planes through a probability calculation module ;
S4, designing a context-based rate reduction module (Rate Reduction Model, RRM), predicting the value of each tri-plane and fusing the values with a convolution layer through a residual block, and finally determining by using an activation functionA value;
S5, designing a context-based distortion reduction module (Distortion Reduction Model, DRM) and using the bias hidden tensor to perform random after entropy decoding Reconstructing an image by using values in a range;
and S6, finally, reconstructing the improved potential tensor image through a super decoder.
By adopting an MPEG video stream image compression algorithm, good video quality under high compression ratio and low bit rate is realized through motion compensation, discrete Cosine Transform (DCT) and Huffman coding technology, and particularly, the method is good at processing motion scenes and reducing time redundancy; in addition, the method overcomes the defects of other series of algorithms in the aspects of dynamic video compression efficiency and bandwidth adaptability, and provides a more flexible and efficient video transmission solution.
Further, S2 includes the following steps:
The encoder extracts the key features in the image, and finally outputs a potential tensor Y by using the progressive extraction of the image features layer by layer of the convolutional neural network; the super-coder further refines and abstracts the features, further compresses the information, etc., to obtain the super-potential tensor Z.
Furthermore, byCalculation of quantized super potential tensorThe average value M and standard deviation are then obtained using a super decoder, as shown in equation (1);
(1)。
Further, S3 includes the following steps:
Using the mean, standard deviation and encoded tri-plane by the probability calculation module WhereinC represents the number of channels, H and W represent the height and width, respectively;
Using entropy parameter M and standard deviation And the coded trihedral value is used for evaluationWherein entropy parameters are typically used to adjust and control the entropy of information during encoding, thereby affecting the loss of information in compression rate.
Further, S4 includes the following steps:
by designing a context-based rate reduction module (Rate Reduction Model, RRM), predicting potential elements in raster scan order is eliminated, but each plane is predicted ,In addition, RRM module pairRefining the probability estimates in (2) to produce updated tensors,Bit ratio required in entropy codingLess, thereby improving the performance of the "rate-distortion" RD.
As shown in fig. 2, RRM is input from the input respectivelyAnd extracting features in context and fusing them by a residual block convolution layer, wherein the fusion tensor has a value equal to that of the residual block convolution layerThe same spatial resolution, but the channels areFour times that of an additively added item by channelAnd a scale item;
First, S is converted to B as shown in equation 2:
(2);
specifically, each element thereof is in In (C) and then willAdded toAnd sums it to produce an updated probability tensor by B modulation。
Further, S5 includes the following steps:
The context-based distortion reduction module (Distortion Reduction Model, DRM) is designed to be used after entropy decoding, unlike non-progressive encoders, where the algorithm proposed herein can use a bias hidden tensor For any one ofReconstructing an imageThus, in decodingThe DRM module then uses the context to reduce errors, thereby reducing image distortion.
As shown in fig. 3, the architecture of DRM is shown. By M andFor context, the DRM module shifts the tensorIs to be ofAs shown in formula (3):
(3);
Unlike RRM, DRM does not use AndAs a context, as it includesDue to the probability information of (2)Has been decoded and used for reconstruction,AndHardly provideIn addition, the DRM module is a regressor for reducing distortion, and the RRM is a classifier for reducing bit rate.
The DRM module is trained to minimize the loss as shown in equation (4):
(4);
finally, the improved latent tensor is passed through the super decoder Image reconstruction is carried out to obtain。
As shown in fig. 4, a terminal device includes:
The system comprises a processor, a memory and a program or instructions stored on the memory and capable of running on the processor, wherein the processor comprises one or more than one processing core, and the processor is connected with the memory through a bus; the processor implements the image compression algorithm described above when executing a program or instructions.
The memory may be implemented by any type of volatile or nonvolatile memory device or combination such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disk.
In addition, the invention also provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program realizes the steps of the context-based progressive three-plane coding image compression algorithm when being executed by a processor. Optionally, the present invention also provides a computer program product comprising instructions which, when run on a computer, cause the computer to perform the steps of a context-based progressive three-plane coded image compression algorithm of the above aspects.
It will be appreciated by those of ordinary skill in the art that the processes for implementing all or part of the steps of the above embodiments may be implemented by hardware, or may be implemented by a program for instructing the relevant hardware, and the program may be stored in a computer readable storage medium, where the above storage medium may be a read-only memory, a magnetic disk or optical disk, etc.
Claims (2)
1. A context-based progressive three-plane coded image compression algorithm, comprising the steps of:
s1, compressing an original video stream by adopting an MPEG algorithm, and obtaining a visual image coded into an H265 format;
s2, sequentially converting the visual image X into a potential tensor Y and a super potential tensor Z through an encoder and a super encoder, and using quantization The super decoder obtains the average value and standard deviation of the representation Y;
S3, evaluating the value by using the average value, the standard deviation and the coded three planes through a probability calculation module ;
S4, designing a context-based rate reduction module, predicting the value of each tri-plane, fusing the values through a residual block and a convolution layer, and finally determining by using an activation functionA value;
s5, designing a context-based distortion reduction module, and using the bias hidden tensor to perform random after entropy decoding Reconstructing an image by using values in a range;
s6, finally, reconstructing the improved potential tensor through a super decoder;
the step S1 comprises the following steps:
an MPEG video stream image compression algorithm is adopted, and video quality under high compression ratio and low bit rate is improved through motion compensation, discrete Cosine Transform (DCT) and Huffman coding technology;
The step S2 comprises the following steps:
the encoder extracts the key features in the image by using a convolutional neural network layer by layer, and finally outputs a potential tensor Y; the super encoder further refines and abstracts the features, further compresses information and the like to obtain a super potential tensor Z;
By passing through Calculation of quantized super potential tensorThe average value M and standard deviation are then obtained using a super decoder, as shown in equation (1);
(1);
The step S3 comprises the following steps:
Using the mean, standard deviation and encoded tri-plane by the probability calculation module WhereinC represents the number of channels, H and W represent the height and width, respectively;
Using entropy parameter M and standard deviation And the coded trihedral value is used for evaluationWherein entropy parameters are typically used to adjust and control the entropy of information during encoding, thereby affecting the loss of information in compression rate;
the step S4 comprises the following steps:
By designing a context-based rate reduction module, predicting potential elements in raster scan order is eliminated, but each plane is predicted ,; In addition, the rate reduction module pairRefining the probability estimates in (2) to produce updated tensors,Bit ratio required in entropy codingLess, thereby improving the performance of the "rate-distortion" RD;
the rate reduction modules respectively receive the input And extracting features in context and fusing them by a residual block convolution layer, wherein the fusion tensor has a value equal to that of the residual block convolution layerThe same spatial resolution, but the channels areFour times that of an additively added item by channelAnd a scale item;
First, S is converted to B as shown in formula (2):
(2);
specifically, each element thereof is in In (C) and then willAdded toAnd sums it to produce an updated probability tensor by B modulation;
The step S5 comprises the following steps:
Designing a context-based distortion reduction module using a bias hidden tensor For any one ofReconstructing an imageThus, in decodingThen, the distortion reduction module uses the context to reduce errors, thereby reducing image distortion;
architecture of the distortion reduction module is M sum For context, the distortion reduction module shifts the tensorIs to be ofAs shown in formula (3):
(3);
The distortion reduction module is trained to minimize losses as shown in equation (4):
(4);
the step S6 comprises the following steps:
Improved latent tensor by super decoder Image reconstruction is carried out to obtain。
2. A terminal device, comprising:
A processor, a memory, and a program or instructions stored on the memory and executable on the processor; the processor, when executing the program or instructions, implements the image compression algorithm of claim 1.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410879262.3A CN118413675B (en) | 2024-07-02 | 2024-07-02 | Context-based progressive three-plane coding image compression algorithm and terminal equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410879262.3A CN118413675B (en) | 2024-07-02 | 2024-07-02 | Context-based progressive three-plane coding image compression algorithm and terminal equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN118413675A CN118413675A (en) | 2024-07-30 |
CN118413675B true CN118413675B (en) | 2024-09-24 |
Family
ID=91990119
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202410879262.3A Active CN118413675B (en) | 2024-07-02 | 2024-07-02 | Context-based progressive three-plane coding image compression algorithm and terminal equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN118413675B (en) |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114008997A (en) * | 2019-06-28 | 2022-02-01 | 黑莓有限公司 | Context determination of planar modes in octree-based point cloud coding |
CN115457182A (en) * | 2022-09-28 | 2022-12-09 | 南京邮电大学 | Interactive viewpoint image synthesis method based on multi-plane image scene representation |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2021255605A1 (en) * | 2020-06-19 | 2021-12-23 | Nokia Technologies Oy | Apparatus, method and computer program product for optimizing parameters of a compressed representation of a neural network |
GB202111188D0 (en) * | 2021-08-03 | 2021-09-15 | Deep Render Ltd | Dr big book 5 |
-
2024
- 2024-07-02 CN CN202410879262.3A patent/CN118413675B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114008997A (en) * | 2019-06-28 | 2022-02-01 | 黑莓有限公司 | Context determination of planar modes in octree-based point cloud coding |
CN115457182A (en) * | 2022-09-28 | 2022-12-09 | 南京邮电大学 | Interactive viewpoint image synthesis method based on multi-plane image scene representation |
Also Published As
Publication number | Publication date |
---|---|
CN118413675A (en) | 2024-07-30 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112866694B (en) | Intelligent image compression optimization method combining asymmetric convolution block and condition context | |
US8295342B2 (en) | Method and system for efficient video compression with low-complexity encoder | |
US11012718B2 (en) | Systems and methods for generating a latent space residual | |
CN108174218B (en) | Video coding and decoding system based on learning | |
CN111314709A (en) | Video compression based on machine learning | |
CN115668777A (en) | System and method for improved machine learning compression | |
KR102245682B1 (en) | Apparatus for compressing image, learning apparatus and method thereof | |
CN116916036A (en) | Video compression method, device and system | |
CN118413675B (en) | Context-based progressive three-plane coding image compression algorithm and terminal equipment | |
US20240163485A1 (en) | Multi-distribution entropy modeling of latent features in image and video coding using neural networks | |
CN110730347A (en) | Image compression method and device and electronic equipment | |
WO2023082107A1 (en) | Decoding method, encoding method, decoder, encoder, and encoding and decoding system | |
CN110717948A (en) | Image post-processing method, system and terminal equipment | |
CN117915107B (en) | Image compression system, image compression method, storage medium and chip | |
US11979587B2 (en) | Hybrid inter-frame coding using an autoregressive model | |
US20240020884A1 (en) | Online meta learning for meta-controlled sr in image and video compression | |
WO2024169959A1 (en) | Method, apparatus, and medium for visual data processing | |
US20080240251A1 (en) | Method For the Encoding of Wavelet-Encoded Images With Bit Rate Control, Corresponding Encoding Device and Computer Program | |
WO2024140849A1 (en) | Method, apparatus, and medium for visual data processing | |
WO2024169958A1 (en) | Method, apparatus, and medium for visual data processing | |
US20220321879A1 (en) | Processing image data | |
WO2024193709A1 (en) | Method, apparatus, and medium for visual data processing | |
WO2024149392A1 (en) | Method, apparatus, and medium for visual data processing | |
CN109803147B (en) | Transformation processing method and device based on video texture features | |
KR20240137005A (en) | Data processing methods, devices and media |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |