CN111800633A - Image processing method and device - Google Patents
Image processing method and device Download PDFInfo
- Publication number
- CN111800633A CN111800633A CN202010581960.7A CN202010581960A CN111800633A CN 111800633 A CN111800633 A CN 111800633A CN 202010581960 A CN202010581960 A CN 202010581960A CN 111800633 A CN111800633 A CN 111800633A
- Authority
- CN
- China
- Prior art keywords
- macro block
- image
- data
- macroblock
- frame image
- 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.)
- Pending
Links
- 238000003672 processing method Methods 0.000 title claims abstract description 39
- 238000007781 pre-processing Methods 0.000 claims abstract description 37
- 238000012545 processing Methods 0.000 claims abstract description 30
- 230000005540 biological transmission Effects 0.000 claims abstract description 23
- 238000013075 data extraction Methods 0.000 claims abstract description 19
- 238000000034 method Methods 0.000 claims description 40
- 230000009466 transformation Effects 0.000 claims description 7
- 238000013527 convolutional neural network Methods 0.000 claims description 4
- 238000007906 compression Methods 0.000 abstract description 10
- 230000006835 compression Effects 0.000 abstract description 10
- 238000010586 diagram Methods 0.000 description 14
- 238000012549 training Methods 0.000 description 11
- 230000000694 effects Effects 0.000 description 8
- 238000013507 mapping Methods 0.000 description 6
- 230000000903 blocking effect Effects 0.000 description 5
- 238000013139 quantization Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 101100161928 Solanum tuberosum AC58 gene Proteins 0.000 description 2
- 230000003321 amplification Effects 0.000 description 2
- 230000006854 communication Effects 0.000 description 2
- 238000013500 data storage Methods 0.000 description 2
- 238000003199 nucleic acid amplification method Methods 0.000 description 2
- 230000016776 visual perception Effects 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013144 data compression Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 239000013598 vector Substances 0.000 description 1
Images
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
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/17—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
- H04N19/176—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object the region being a block, e.g. a macroblock
-
- 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
- H04N19/124—Quantisation
-
- 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
- H04N19/42—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation
-
- 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
- H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
- H04N19/625—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding using discrete cosine transform [DCT]
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Discrete Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The disclosure provides an image processing method and an image processing device, relates to the field of image processing, and can solve the problem that data volume in an image transmission process cannot be reduced to the maximum extent by image compression. The specific technical scheme is as follows: acquiring a frame image to be processed, and performing macro block division on the frame image to be processed to generate at least one macro block; preprocessing each macro block in the at least one macro block to obtain a preprocessed macro block corresponding to each macro block; layering each preprocessed macro block according to a preset layering rule to obtain layered data corresponding to each preprocessed macro block; extracting the hierarchical data corresponding to each preprocessed macro block according to a preset data extraction rule to obtain target data corresponding to each preprocessed macro block; and sending the target data corresponding to each preprocessed macro block to image receiving equipment. The invention is used for reducing the data volume in the image transmission process.
Description
Technical Field
The present disclosure relates to the field of image processing, and in particular, to an image processing method and apparatus.
Background
With the development of information technology, the requirements of people on communication services are continuously increased, and more images, especially video images, are transmitted in the communication process. In the process of transmitting the video image, the data size is huge, and if the video image is not compressed, the video image is difficult to transmit and store. Therefore, image compression capability is particularly important for video images.
Image compression is the application of data compression techniques to digital images with the goal of reducing redundant information in the image data for more efficient format storage and transmission of the image.
However, the current image compression generally compresses the whole data of the image, and this way cannot reduce the data amount in the image transmission process to the maximum extent.
Disclosure of Invention
The embodiment of the disclosure provides an image processing method and an image processing device, which can solve the problem that the data volume in the image transmission process cannot be reduced to the maximum extent by the current image compression. The technical scheme is as follows:
according to a first aspect of the embodiments of the present disclosure, there is provided an image processing method applied to an image transmission apparatus, the method including:
acquiring a frame image to be processed, and performing macro block division on the frame image to be processed to generate at least one macro block;
preprocessing each macro block in the at least one macro block to obtain a preprocessed macro block corresponding to each macro block, wherein each preprocessed macro block comprises a Direct Current (DC) component and at least one Alternating Current (AC) component;
layering each preprocessed macro block according to a preset layering rule to obtain layered data corresponding to each preprocessed macro block, wherein the layered data comprises a base layer and at least one enhancement layer, the DC component is the base layer, the AC component is divided into at least one enhancement layer according to importance degrees, and the importance degrees of the AC components in the same enhancement layer are the same;
extracting the hierarchical data corresponding to each preprocessed macro block according to a preset data extraction rule to obtain target data corresponding to each preprocessed macro block, wherein the target data comprises a base layer data and at least one enhancement layer data;
and sending the target data corresponding to each preprocessed macro block to image receiving equipment.
The image processing method provided by the embodiment of the disclosure can acquire a frame image to be processed, and perform macro block division on the frame image to be processed to generate at least one macro block; preprocessing each macro block in the at least one macro block to obtain a preprocessed macro block corresponding to each macro block, wherein each preprocessed macro block comprises a DC component and at least one AC component; extracting the hierarchical data corresponding to each preprocessed macro block according to a preset data extraction rule to obtain target data corresponding to each preprocessed macro block, wherein the target data comprises a basic layer data and at least one enhancement layer data; and sending the target data corresponding to each pre-processing macro block to image receiving equipment, and only sending the target data corresponding to each pre-processing macro block in the image transmission process, wherein the transmitted data volume is less, and the problem that the data volume in the image transmission process cannot be reduced to the maximum extent by the conventional image compression is solved.
In one embodiment, the target data includes one base layer data and three enhancement layer data, the three enhancement layer data being data of three enhancement layers of highest importance in the AC component.
The target data comprises a base layer data and three enhancement layer data, only one base layer data and three enhancement layer data corresponding to each pre-processed macro block need to be sent to the image receiving equipment, only the target data corresponding to each pre-processed macro block is sent in the image transmission process, the transmitted data volume is small, and the problem that the data volume in the image transmission process cannot be reduced to the maximum degree by the existing image compression data is solved.
In one embodiment, the pre-processing each of the at least one macroblock comprises:
performing DCT (discrete cosine transformation) on each macro block in the at least one macro block to obtain a transformed macro block corresponding to each macro block, wherein the transformed macro block comprises a DC component and at least one AC component;
and quantizing each transformed macroblock to obtain a quantized macroblock corresponding to each transformed macroblock, wherein the quantized macroblock comprises a quantized DC component and at least one quantized AC component.
By performing DCT transformation and quantization processing on each macro block of the at least one macro block, the data processing amount can be effectively reduced.
In one embodiment, before the target data corresponding to each pre-processed macroblock is sent to an image receiving device, the method further includes:
and coding the target data corresponding to each pre-processed macro block to obtain a coded macro block corresponding to each pre-processed macro block.
By encoding the target data corresponding to each preprocessed macro block, the safety of the target data in the transmission process can be improved.
According to a second aspect of the embodiments of the present disclosure, there is provided an image processing method applied to an image receiving apparatus, the method including:
receiving target data of each macro block in the first aspect sent by an image sending device;
reducing the dimension of each macro block according to the gray level of the basic layer data of each macro block to obtain a frame image after dimension reduction;
amplifying the reduced frame image into a frame image with a target size;
and performing super-resolution reconstruction on the frame image with the target size and the at least one enhancement layer data to generate a target frame image.
The image processing method provided by the embodiment of the disclosure can receive target data of each macro block sent by an image sending device; reducing the dimension of each macro block according to the gray level of the basic layer data of each macro block to obtain a frame image after dimension reduction; amplifying the reduced frame image into a frame image with a target size; the frame image with the target size and the at least one enhancement layer data are subjected to super-resolution reconstruction to generate a target frame image, only target data corresponding to each pre-processed macro block are received in the image transmission process, the transmitted data volume is small, the problem that the data volume in the image transmission process cannot be reduced to the maximum extent in the conventional image compression is solved, and the image processing method provided by the disclosed embodiment reduces the target frame image block effect and improves the resolution of the target frame image.
In one embodiment, the super-resolution reconstruction of the target size frame image and the at least one enhancement layer data comprises:
and inputting the frame image with the target size and the at least one enhancement layer data into a pre-trained super-resolution reconstruction convolutional neural network (SRCNN) model.
The frame image with the target size and the data of the at least one enhancement layer are input into a pre-trained super-resolution reconstruction convolutional neural network (SRCNN) model, so that the resolution of the target frame image can be improved.
In one embodiment, each macro block sent by the image sending device is a coded macro block, and after receiving target data of each macro block sent by the image sending device, the method further includes:
and decoding each macro block to generate a decoded macro block corresponding to each macro block.
By decoding each macroblock, target data of the macroblock can be efficiently recovered.
According to a third aspect of the embodiments of the present disclosure, there is provided an image processing apparatus applied to an image transmission device, the apparatus including:
the frame image to be processed acquiring module is used for acquiring a frame image to be processed and dividing macro blocks of the frame image to be processed to generate at least one macro block;
a macroblock preprocessing module, configured to preprocess each macroblock of the at least one macroblock to obtain a preprocessed macroblock corresponding to each macroblock, where each preprocessed macroblock includes a DC component and at least one AC component;
the device comprises a preprocessing macro block layering module, a preprocessing macro block layering module and a control module, wherein the preprocessing macro block layering module is used for layering each preprocessing macro block according to a preset layering rule to obtain layering data corresponding to each preprocessing macro block, the layering data comprise a base layer and at least one enhancement layer, the DC component is the base layer, the AC component is divided into at least one enhancement layer according to importance degrees, and the importance degrees of the AC components in the same enhancement layer are the same;
the hierarchical data extraction module is used for extracting the hierarchical data corresponding to each preprocessed macro block according to a preset data extraction rule to obtain target data corresponding to each preprocessed macro block, wherein the target data comprises a base layer data and at least one enhancement layer data;
and the target data sending module is used for sending the target data corresponding to each preprocessed macro block to the image receiving equipment.
In one embodiment, the target data includes one base layer data and three enhancement layer data, the three enhancement layer data being data of three enhancement layers of highest importance in the AC component.
In one embodiment, the macroblock preprocessing module is specifically configured to:
performing DCT (discrete cosine transformation) on each macro block in the at least one macro block to obtain a transformed macro block corresponding to each macro block, wherein the transformed macro block comprises a DC component and at least one AC component;
and quantizing each transformed macroblock to obtain a quantized macroblock corresponding to each transformed macroblock, wherein the quantized macroblock comprises a quantized DC component and at least one quantized AC component.
In one embodiment, the apparatus further comprises:
and the macro block coding module is used for coding the target data corresponding to each pre-processed macro block to obtain a coded macro block corresponding to each pre-processed macro block.
According to a fourth aspect of the embodiments of the present disclosure, there is provided an image processing apparatus applied to an image receiving device, the method including:
a target data receiving module, configured to receive target data of each macro block according to the first aspect, sent by an image sending device;
the macro block dimension reduction module is used for reducing the dimension of each macro block according to the gray level of the basic layer data of each macro block to obtain a frame image after dimension reduction;
the frame image amplifying module is used for amplifying the frame image subjected to the dimensionality reduction into a frame image with a target size;
and the target frame image generation module is used for performing super-resolution reconstruction on the frame image with the target size and the at least one enhancement layer data to generate a target frame image.
In one embodiment, the apparatus further comprises:
and the macro block decoding module is used for decoding each macro block and generating a decoded macro block corresponding to each macro block.
A fifth aspect of embodiments of the present disclosure provides an image processing apparatus comprising a processor and a memory, the memory having stored therein at least one computer instruction, the instruction being loaded and executed by the processor to implement the steps performed in the image processing method of any one of the first aspects.
A sixth aspect of the embodiments of the present disclosure provides a computer-readable storage medium, having stored therein at least one computer instruction, which is loaded and executed by a processor to implement the steps performed in the image processing method according to any one of the first aspects.
A seventh aspect of the embodiments of the present disclosure provides an image processing apparatus, which includes a processor and a memory, where at least one computer instruction is stored in the memory, and the instruction is loaded and executed by the processor to implement the steps performed in the image processing method according to any one of the second aspects.
An eighth aspect of the embodiments of the present disclosure provides a computer-readable storage medium, in which at least one computer instruction is stored, the instruction being loaded and executed by a processor to implement the steps performed in the image processing method according to any one of the second aspects.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and together with the description, serve to explain the principles of the disclosure.
FIG. 1 is a first flowchart of an image processing method provided by the disclosed embodiments;
fig. 2 is a schematic diagram illustrating the hierarchical rule of an 8 × 8 macroblock according to an embodiment of the disclosure;
fig. 3 is a flowchart ii of an image processing method provided in the embodiment of the present disclosure;
fig. 4 is a flowchart three of an image processing method provided by the embodiment of the present disclosure;
fig. 5 is a fourth flowchart of an image processing method provided by the embodiment of the present disclosure;
fig. 6 is a structural diagram of an SRCNN model provided in an embodiment of the present disclosure;
fig. 7 is a parameter structure diagram of an srnnn model provided in an embodiment of the present disclosure;
fig. 8 is a flowchart of a method for training an srnnn model according to an embodiment of the present disclosure;
fig. 9 is a first structural diagram of an image processing apparatus according to an embodiment of the present disclosure;
FIG. 10 is a block diagram of an image processing apparatus according to the second embodiment;
FIG. 11 is a block diagram of an image processing apparatus according to a third embodiment of the disclosure;
FIG. 12 is a block diagram of an image processing apparatus according to a fourth embodiment of the disclosure;
FIG. 13 is a block diagram of an image processing apparatus provided in the disclosed embodiments;
fig. 14 is a block diagram of an image processing apparatus according to a second embodiment of the disclosure.
Detailed Description
Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, like numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the exemplary embodiments below are not intended to represent all implementations consistent with the present disclosure. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present disclosure, as detailed in the appended claims.
Fig. 1 is a first flowchart of an image processing method provided by an embodiment of the present disclosure, where the image processing method is applied to an image sending device, which is an image encoding end. As shown in fig. 1, the image processing method includes the steps of:
s101, obtaining a frame image to be processed, and performing macro block division on the frame image to be processed to generate at least one macro block.
Illustratively, the at least one macroblock may be a 4 × 4 macroblock, an 8 × 8 macroblock, a 16 × 16 macroblock, or other size macroblocks. Preferably, in the present embodiment, the at least one macroblock is an 8 × 8 macroblock.
S102, each macro block in the at least one macro block is preprocessed to obtain a preprocessed macro block corresponding to each macro block, and each preprocessed macro block comprises a Direct Current (DC) component and at least one Alternating Current (AC) component.
How each of the at least one macroblock is preprocessed is explained as follows:
exemplarily, each macroblock in the at least one macroblock is subjected to Discrete Cosine Transform (DCT) to obtain a transformed macroblock corresponding to each macroblock, where the transformed macroblock includes a DC component and at least one AC component;
and then quantizing each transformed macroblock to obtain a quantized macroblock corresponding to each transformed macroblock, wherein the quantized macroblock comprises a quantized DC component and at least one quantized AC component.
Illustratively, a pre-processed macroblock resulting from pre-processing an 8 × 8 macroblock includes one DC component and 63 AC components.
S103, layering each preprocessed macro block according to a preset layering rule to obtain layered data corresponding to each preprocessed macro block, wherein the layered data comprises a base layer and at least one enhancement layer, the DC component is the base layer, the AC component is divided into at least one enhancement layer according to the importance degree, and the importance degrees of the AC components in the same enhancement layer are the same.
Illustratively, the preset layering rule specifies the number of layers to be divided for each macroblock, and the positions of the pixels included in each layer in the macroblock. In practical applications, for 8 × 8 macroblocks, 16 layers may be divided.
Fig. 2 is a schematic diagram of a hierarchical rule of an 8 × 8 macroblock shown in the embodiment of the present disclosure, and referring to fig. 2, pixels belonging to the same layer are labeled with the same color, for example, a DC component is a first layer base layer, AC components such as AC1 and AC2 belong to a second layer enhancement layer, AC3 and AC4 belong to a third layer enhancement layer, AC5, AC6 and AC7 belong to a fourth layer enhancement layer, and so on, AC58, AC59, AC60, AC61, AC62 and AC63 belong to a sixteenth layer enhancement, and so on, where the DC component has the highest importance level, and the AC components have the highest importance level in the order of the second layer enhancement layer and the third enhancement layer … … from high to low, and up to the sixteenth enhancement layer, and the AC components in the same enhancement layer have the same importance level.
In practical application, the layering rules can be set and adjusted according to actual needs, and once the layering rules are determined, all macro blocks in the current frame image are layered according to the layering rules, so that pixel points contained in each layer are determined.
And S104, extracting the layered data corresponding to each preprocessed macro block according to a preset data extraction rule to obtain target data corresponding to each preprocessed macro block, wherein the target data comprises a base layer data and at least one enhancement layer data.
The predetermined data extraction rule requires specifying the number of data layers to be extracted for each pre-processed macroblock. Generally, lower layer data is selected preferentially. For example, the first layer may be selected, that is, the base layer data is transmitted, or the 1 st to N th layers may be selected, and the value of N may be set according to actual needs. And according to the empirical value, the value of N is not more than 4.
In this embodiment, N is 4, i.e., the target data includes one base layer data and three enhancement layer data, which are data of three enhancement layers of the highest importance in the AC component. Taking an 8 × 8 macroblock as an example, the target data includes first base layer data and a second enhancement layer, a third enhancement layer, and a fourth enhancement layer data.
And S105, sending the target data corresponding to each preprocessed macro block to the image receiving equipment.
Illustratively, the target data corresponding to each pre-processed macro block is encoded to obtain an encoded macro block corresponding to each pre-processed macro block, and then the encoded macro block corresponding to each pre-processed macro block is sent to the image receiving device.
The image processing method provided by the embodiment of the disclosure can acquire a frame image to be processed, and perform macro block division on the frame image to be processed to generate at least one macro block; preprocessing each macro block in the at least one macro block to obtain a preprocessed macro block corresponding to each macro block, wherein each preprocessed macro block comprises a DC component and at least one AC component; extracting the hierarchical data corresponding to each preprocessed macro block according to a preset data extraction rule to obtain target data corresponding to each preprocessed macro block, wherein the target data comprises a basic layer data and at least one enhancement layer data; and sending the target data corresponding to each pre-processing macro block to image receiving equipment, and only sending the target data corresponding to each pre-processing macro block in the image transmission process, wherein the transmitted data volume is less, and the problem that the data volume in the image transmission process cannot be reduced to the maximum extent by the conventional image compression is solved.
The following describes the image processing method shown in the embodiment of fig. 1 in further detail with reference to the embodiment of fig. 3. Fig. 3 is a flowchart ii of an image processing method according to an embodiment of the present disclosure. As shown in fig. 3, the image processing method includes the steps of:
s301, acquiring a current frame image;
s302, preprocessing the current frame image, wherein the preprocessing comprises the following steps: DCT transformation and quantization;
specifically, the preprocessing of the current frame image is performed in units of macroblocks, that is, before the preprocessing, the entire current frame image is divided into a plurality of macroblocks, and then the preprocessing is performed on each macroblock, the size of each macroblock can be set according to actual needs, and in JPEG encoding, the image frame is usually divided into 8 × 8 macroblocks for processing.
The DCT transform can convert the signals in the spatial domain to the frequency domain, and has good decorrelation performance. The DCT is lossless, but creates good conditions for the next quantization, coding and the like in the image coding process, and meanwhile, because the DCT is symmetrical, the original image information can be restored by using the DCT inverse transformation at the receiving end after the quantization coding.
The data obtained by preprocessing the current frame image mainly includes a DC component part and an AC component part, and specifically, for an 8 × 8 macroblock, one DC component and 63 AC components are obtained.
S303, carrying out layering processing on the preprocessed frame image;
the layering is still performed in units of macro blocks, and specifically, the layering is performed according to a preset layering rule for each macro block. The preset layering rule specifies the number of layers divided by each macro block and the positions of pixel points contained in each layer in the macro block. In practical applications, for 8 × 8 macroblocks, 16 layers may be divided.
Specifically, referring to fig. 2, the pixels belonging to the same layer are labeled with the same color, for example, DC is the first layer base layer, AC1 and AC2 belong to the second layer enhancement layer, AC3 and AC4 belong to the third layer enhancement layer, AC5, AC6 and AC7 belong to the fourth layer enhancement layer, and so on, AC58, AC59, AC60, AC61, AC62 and AC63 belong to the sixteenth layer enhancement, and so on.
In practical application, the layering rules can be set and adjusted according to actual needs, and once the layering rules are determined, all macro blocks in the current frame image are layered according to the layering rules, so that pixel points contained in each layer are determined.
The hierarchical rules are set based on pre-processed data, which generally includes a DC component part and an AC component part, and the basic principle is as follows: taking the DC component as a first layer and mainly taking charge of displaying the outline part of the whole picture; and layering the AC components according to the importance degree of the AC components, wherein the importance degree is divided into the same layer. After the layers are sorted, the layers are numbered, and the importance level is gradually decreased from the layer 2 to the layer 16, which is the layer with the higher importance level.
S304, taking out one or more layers of preset data from the current frame image;
in this step, one or more layers of data are extracted according to a preset data extraction rule.
The data extraction rules need to specify the number of data layers to be extracted. Generally, lower layer data is preferentially selected for transmission, for example, layer 1 may be selected, that is, base layer data is transmitted, or layers 1 to N may be selected, and the value of N may be set according to actual needs. And according to the empirical value, the value of N is not more than 4.
Once the data extraction rule is determined, one or more layers of data required for each macro block in the current image frame are extracted according to the same data extraction rule, so that data extraction of the whole current frame image is completed.
S305, encoding the extracted data, and storing the encoded data or transmitting the encoded data to an image receiving apparatus through a network.
In this step, the extracted data may be encoded according to a preset encoding method, where the preset encoding method may be any one or more existing encoding methods.
How the image receiving apparatus performs image processing will be described with reference to the embodiment of fig. 4. Fig. 4 is a flowchart of a third method for processing an image according to an embodiment of the present disclosure, where the method is applied to an image receiving device, and the image receiving device is an image decoding end. As shown in fig. 4, the image processing method includes the steps of:
s401, receiving target data of each macro block sent by the image sending device.
Illustratively, the target data includes a base layer data and at least one enhancement layer data. Taking an 8 × 8 macroblock as an example, the target data includes first base layer data and a second enhancement layer, a third enhancement layer, and a fourth enhancement layer data.
In this embodiment, each macroblock sent by the image sending device is a coded macroblock, and after receiving target data of each macroblock sent by the image sending device, the image receiving device needs to decode each macroblock to generate a decoded macroblock corresponding to each macroblock.
S402, reducing the dimension of each macro block according to the gray level of the basic layer data of each macro block to obtain a reduced-dimension frame image;
and S403, amplifying the frame image after the dimension reduction into a frame image with a target size.
The most obvious blocking effect in an image is the base layer data in the macro blocks, in the base layer data, because each divided 8 × 8 macro block only has a low frequency component, that is, the shadow intensity information of the image, and therefore each 8 × 8 macro block has the same gray value, each 8 × 8 macro block can be reduced according to the gray level of the base layer data of the 8 × 8 macro block, for example, a frame image of 1920 × 1080 is divided into 8 × 8 macro blocks, and then the frame image of 1920 × 1080 is reduced according to the gray level of the base layer data of each macro block to obtain a reduced frame image, which can be represented as a 240 × 135 matrix, and the reduced frame image is a low-resolution image.
The frame image after dimensionality reduction is enlarged into a frame image with a target size by interpolation amplification, for example, bicubic interpolation, for example, the frame image after dimensionality reduction of a 240 × 135 matrix is enlarged into a frame image of 1920 × 1080 by interpolation. The frame image of the target size is referred to as a low resolution image although enlarged by interpolation.
S404, performing super-resolution reconstruction on the frame image with the target size and the at least one enhancement layer data to generate a target frame image.
For example, the target frame image may be generated by super-resolution reconstructing the frame image of the target size and the at least one enhancement layer data using an interpolation-based super-resolution reconstruction algorithm, a reconstruction-based super-resolution reconstruction algorithm, a learning-based super-resolution reconstruction algorithm, or another super-resolution reconstruction algorithm.
In this embodiment, a learning-based super-resolution reconstruction algorithm is used to perform super-resolution reconstruction on the frame image of the target size and the at least one enhancement layer data. Illustratively, the learning-based hyper-Resolution reconstruction algorithm may employ a Super-Resolution reconstruction Convolutional Neural Network (SRCNN) model based on images.
In this embodiment, the frame image of the target size and the data of the at least one enhancement layer are input into a pre-trained srnnn neural network model, so as to generate a target frame image, which is a high-resolution frame image.
As the blocking effect of the data of the 2 nd-4 th layers of enhancement layers in the image is not obvious, the data of the 2 nd-4 th layers of enhancement layers are directly input into the SRCNN model without the processes of dimension reduction and interpolation amplification, and the high-resolution image with the super-resolution deblocking effect can be reconstructed.
For enhancement layer data larger than layer 4, the superposition of the base layer data and the enhancement layer data makes the enhancement layer data larger than layer 4 have almost no blocking effect in the image, and does not affect the visual perception of the user even if the enhancement layer data larger than layer 4 is not processed, so in this application, the image transmitting apparatus does not transmit the enhancement layer data larger than layer 4 to the image receiving apparatus, i.e., the enhancement layer data larger than layer 4 is not included in the target data of each macroblock received by the image receiving apparatus.
The image processing method provided by the embodiment of the disclosure can receive target data of each macro block sent by an image sending device; reducing the dimension of each macro block according to the gray level of the basic layer data of each macro block to obtain a frame image after dimension reduction; amplifying the reduced frame image into a frame image with a target size; the frame image with the target size and the at least one enhancement layer data are subjected to super-resolution reconstruction to generate a target frame image, only target data corresponding to each pre-processed macro block are received in the image transmission process, the transmitted data volume is small, the problem that the data volume in the image transmission process cannot be reduced to the maximum extent in the conventional image compression is solved, and the image processing method provided by the disclosed embodiment reduces the target frame image block effect and improves the resolution of the target frame image.
The image processing method provided in the embodiment of fig. 4 will be further described in detail with reference to the embodiment of fig. 5. Fig. 5 is a fourth flowchart of an image processing method provided by an embodiment of the present disclosure, where the method is applied to an image receiving device, which is an image decoding end. As shown in fig. 5, the image processing method includes the steps of:
s501, receiving frame image data sent by an image coding end;
s502, decoding the received frame image data to obtain one layer or multiple layers of image data;
in this step, the currently received code stream is decoded in a decoding manner corresponding to the encoding manner adopted by the image encoding end. The one or more layers of image data are one base layer data per macroblock or one base layer data and at least one enhancement layer data per macroblock.
And S503, performing super-resolution reconstruction on the one or more layers of image data obtained in the step S502 by adopting an SRCNN model, and obtaining a high-definition image after reconstruction.
How to perform image reconstruction according to the SRCNN model is described below.
By analyzing the layered images, it is found that the blocking effect is most obvious in the first layer image, i.e. the base layer data, in the first layer image data, since each divided 8 × 8 macroblock only has a low frequency component, i.e. shadow intensity information of the image, and therefore, the same gray value exists in each 8 × 8 macroblock, for the first layer image, each 8 × 8 macroblock can be represented by the gray value of the base layer data, for example, a 1920 × 1080 image is represented as a 240 × 135 matrix, and the image is restored to the original image size by the super-resolution algorithm.
There are many existing hyper-resolution methods, such as interpolation-based hyper-resolution reconstruction, reconstruction-based hyper-resolution reconstruction, and learning-based hyper-resolution reconstruction. The learning-based method is to utilize a large amount of training data to learn a certain corresponding relation between a low-resolution image and a high-resolution image, and then predict the high-resolution image corresponding to the low-resolution image according to the learned mapping relation, so as to realize the super-resolution reconstruction process of the image.
According to the scheme, the SRCNN model is adopted to carry out super-resolution reconstruction on the low-resolution image. The SRCNN model is an image hyper-resolution reconstruction based on a deep learning method, and the SRCNN model is adopted to realize end-to-end mapping from low resolution to high resolution images, and the structure of the SRCNN model is shown in FIG. 6.
For a low-resolution image, firstly, a bicubic interpolation is adopted to enlarge the low-resolution image to an image target, then nonlinear mapping is fitted through a three-layer convolution network, and finally, a high-resolution image result is output. The process of reconstructing the image by adopting the SRCNN model comprises the following steps:
step 1: and extracting image blocks. The low resolution image is first scaled to the target size using bicubic interpolation, such as 240 × 135 to 1920 × 1080 using interpolation. Although enlarged by interpolation, it is still called a low resolution image, i.e. a super-resolution input image. And extracting image blocks from the low-resolution input image to form a high-dimensional feature map.
F1(Y)＝max(0，W1*Y+B1)
Wherein, W1And B1Obtained by learning for the over-scoring parameter, F1(Y) is a feature value of the high-dimensional feature map, and Y is a feature value of the enlarged low-resolution image. Illustratively, the high-dimensional feature map can be obtained by inputting the low-resolution image after the method into the first layer of convolution of the SRCNN model. The convolution kernel size of the first layer of convolution is 9 x 9(f1 x f1), the convolution kernel number is 64(n1), and 64 high-dimensional feature maps can be output through the first layer of convolution calculation.
Step 2: and (4) nonlinear mapping. Inputting the high-dimensional feature map output by the first layer of convolution into a second layer of convolution, wherein the convolution kernel size of the second layer of convolution is 1 x 1(f2 x f2), the convolution kernel number is 32(n2), and the second layer of convolution can calculate and output 32 feature maps; this process enables a non-linear mapping of two high-dimensional feature vectors.
F2(Y)＝max(0，W2*F1(Y)+B2)
Wherein, W2N1 x 1 x n 2. Using 1 x 1 convolution, B2As a super-resolution parameter, F2(Y) is the feature value of the 32 feature maps.
Step 3: and (4) reconstructing. And inputting the 32 feature maps output by the second layer of convolution into a third layer of convolution, wherein the convolution kernel of the third layer of convolution is 5 x 5(f3 x f3), the number of the convolution kernels is 1(n3), and 1 feature map is output, namely the final reconstructed high-resolution image.
F3(Y)＝W3*F2(Y)+B3
Wherein, W3And B3As a super-resolution parameter, F3(Y) is the feature value of the final reconstructed high resolution image.
Illustratively, the parameter structure of the srnn model is as shown in fig. 7, and the parameter structure of the srnn model includes:
the first layer of convolution: the convolution kernel size is 9 multiplied by 9(f1 multiplied by f1), the number of convolution kernels is 64(n1), and 64 feature maps are output;
second layer convolution: the convolution kernel size is 1 multiplied by 1(f2 multiplied by f2), the number of convolution kernels is 32(n2), and 32 feature maps are output;
and a third layer of convolution: the convolution kernel size is 5 × 5(f3 × f3), the number of convolution kernels is 1(n3), and 1 feature map is output, namely the final reconstructed high-resolution image.
For the frame images of the 2 nd to 4 th layers, because the blocking effect is not obvious, the process of down sampling and interpolation is not needed, namely the step1 process, the image is directly subjected to step2 nonlinear mapping, and then the image is reconstructed by step3 to obtain the super-resolution de-blocking effect image.
For frame images larger than the layer 4, the superposition of the base layer and the enhancement layer makes the image block effect less obvious, and the visual perception is not affected even if the image block effect is not processed, so that the scheme does not process the frame images transmitted to the layer higher than the layer 4.
How to train the SRCNN model is described below with reference to the embodiment of FIG. 8. Fig. 8 is a flowchart of a method for training an srnnn model according to an embodiment of the present disclosure. As shown in fig. 8, the method includes:
s801, generating a training sample set;
firstly, collecting an image;
for the purpose of training the neural network, a large number of images need to be acquired to generate a training sample set.
Secondly, layering the acquired images;
thirdly, extracting image data of a corresponding layer according to a preset data extraction rule;
fourthly, encoding and decoding the extracted image data;
fifthly, using the image data and the original image obtained after decoding as a training sample set;
wherein, the training set comprises a plurality of training samples, each of which can be expressed as (Ai, Bi), where Ai represents the decoded image data, i.e. the input data; bi represents the original, i.e. the target image.
S802, training the SRCNN model through a training sample set.
Specifically, the image data obtained after decoding is used as the input of the SRCNN model, and the output of the SRCNN model is calculated; calculating the difference between the output of the SRCNN model and the actual value (target), adjusting the parameters (weight) of the SRCNN model according to the obtained difference, and continuously repeating the steps for continuous iteration until the difference between the output of the SRCNN model and the actual value is smaller than a preset error threshold.
And the finally obtained SRCNN model is used as the pre-trained SRCNN model used in the invention and is used for carrying out the super-resolution reconstruction on the low-resolution image at the image decoding end.
Fig. 9 is a first structural diagram of an image processing apparatus provided in an embodiment of the present disclosure, where the apparatus is applied to an image sending device. As shown in fig. 9, the apparatus 90 includes:
a to-be-processed frame image obtaining module 901, configured to obtain a to-be-processed frame image, and perform macroblock division on the to-be-processed frame image to generate at least one macroblock;
a macroblock preprocessing module 902, configured to preprocess each macroblock of the at least one macroblock to obtain a preprocessed macroblock corresponding to each macroblock, where each preprocessed macroblock includes a DC component and at least one AC component;
a pre-processing macroblock layering module 903, configured to layer each pre-processing macroblock according to a preset layering rule to obtain layered data corresponding to each pre-processing macroblock, where the layered data includes a base layer and at least one enhancement layer, the DC component is the base layer, the AC component is divided into at least one enhancement layer according to importance degrees, and the importance degrees of the AC components in the same enhancement layer are the same;
a layered data extraction module 904, configured to extract layered data corresponding to each of the pre-processed macro blocks according to a preset data extraction rule, so as to obtain target data corresponding to each of the pre-processed macro blocks, where the target data includes a base layer data and at least one enhancement layer data;
a target data sending module 905, configured to send the target data corresponding to each preprocessed macro block to an image receiving device.
In one embodiment, the target data includes one base layer data and three enhancement layer data, the three enhancement layer data being data of three enhancement layers of highest importance in the AC component.
In one embodiment, the macroblock preprocessing module 902 is specifically configured to:
performing DCT (discrete cosine transformation) on each macro block in the at least one macro block to obtain a transformed macro block corresponding to each macro block, wherein the transformed macro block comprises a DC component and at least one AC component;
and quantizing each transformed macroblock to obtain a quantized macroblock corresponding to each transformed macroblock, wherein the quantized macroblock comprises a quantized DC component and at least one quantized AC component.
In one embodiment, as shown in fig. 10, the apparatus 90 further comprises:
a macroblock encoding module 906, configured to encode the target data corresponding to each pre-processed macroblock to obtain an encoded macroblock corresponding to each pre-processed macroblock.
For the image processing apparatus provided in the embodiment of the present disclosure, the implementation process and the technical effect thereof can be referred to the embodiments of fig. 1 to fig. 3, and are not described herein again.
Fig. 11 is a third structural diagram of an image processing apparatus provided in an embodiment of the present disclosure, where the apparatus is applied to an image receiving device. As shown in fig. 11, the apparatus 110 includes:
a target data receiving module 1101, configured to receive target data of each macro block according to the first aspect sent by an image sending device;
a macroblock dimension reduction module 1102, configured to perform dimension reduction on each macroblock according to a gray level of the base layer data of each macroblock, to obtain a frame image after dimension reduction;
a frame image enlarging module 1103, configured to enlarge the dimension-reduced frame image into a frame image of a target size;
and a target frame image generating module 1104, configured to perform super-resolution reconstruction on the frame image of the target size and the at least one enhancement layer data, and generate a target frame image.
In one embodiment, as shown in fig. 12, the apparatus 110 further comprises:
a macroblock decoding module 1105, configured to decode each macroblock, and generate a decoded macroblock corresponding to each macroblock.
For the image processing apparatus provided in the embodiment of the present disclosure, the implementation process and the technical effect thereof can be referred to the embodiments of fig. 4 to fig. 7, and are not described herein again.
Fig. 13 is a schematic structural diagram of an image processing apparatus according to an embodiment of the present disclosure, and as shown in fig. 14, the image processing apparatus 130 includes:
a processor 1301 and a memory 1302, the memory 1301 having stored therein at least one computer instruction, the instruction being loaded and executed by the processor 1301 to implement the steps performed in the image processing method described in the embodiments corresponding to fig. 1 to 3.
Based on the image processing methods described in the embodiments corresponding to fig. 1 to fig. 3, embodiments of the present disclosure further provide a computer-readable storage medium, for example, the non-transitory computer-readable storage medium may be a Read Only Memory (ROM), a Random Access Memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like. The storage medium stores computer instructions for executing the image processing method described in the embodiment corresponding to fig. 1 to 3, which is not described herein again.
Fig. 14 is a schematic structural diagram of a second image processing apparatus provided in an embodiment of the present disclosure, and as shown in fig. 14, the image processing apparatus 140 includes:
a processor 1401 and a memory 1402, the memory 1401 having stored therein at least one computer instruction, which instruction is loaded and executed by the processor 1401 to implement the steps performed in the image processing method described in the embodiments corresponding to fig. 4 to 7.
Based on the image processing methods described in the embodiments corresponding to fig. 4 to fig. 7, embodiments of the present disclosure further provide a computer-readable storage medium, for example, the non-transitory computer-readable storage medium may be a Read Only Memory (ROM), a Random Access Memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like. The storage medium stores computer instructions for executing the image processing method described in the embodiment corresponding to fig. 4 to 7, which is not described herein again.
It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program instructing relevant hardware, where the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.
Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This application is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.
Claims (10)
1. An image processing method applied to an image transmission apparatus, the method comprising:
acquiring a frame image to be processed, and performing macro block division on the frame image to be processed to generate at least one macro block;
preprocessing each macro block in the at least one macro block to obtain a preprocessed macro block corresponding to each macro block, wherein each preprocessed macro block comprises a Direct Current (DC) component and at least one Alternating Current (AC) component;
layering each preprocessed macro block according to a preset layering rule to obtain layered data corresponding to each preprocessed macro block, wherein the layered data comprises a base layer and at least one enhancement layer, the DC component is the base layer, the AC component is divided into at least one enhancement layer according to importance degrees, and the importance degrees of the AC components in the same enhancement layer are the same;
extracting the hierarchical data corresponding to each preprocessed macro block according to a preset data extraction rule to obtain target data corresponding to each preprocessed macro block, wherein the target data comprises a base layer data and at least one enhancement layer data;
and sending the target data corresponding to each preprocessed macro block to image receiving equipment.
2. The method of claim 1, wherein the target data comprises one base layer data and three enhancement layer data, and wherein the three enhancement layer data are data of three enhancement layers with the highest significance in the AC component.
3. The method of claim 1, wherein the pre-processing each of the at least one macroblock comprises:
performing Discrete Cosine Transform (DCT) on each macro block in the at least one macro block to obtain a transformed macro block corresponding to each macro block, wherein the transformed macro block comprises a DC component and at least one AC component;
and quantizing each transformed macroblock to obtain a quantized macroblock corresponding to each transformed macroblock, wherein the quantized macroblock comprises a quantized DC component and at least one quantized AC component.
4. The method according to claim 1, wherein before sending the target data corresponding to each pre-processed macro block to an image receiving device, the method further comprises:
and coding the target data corresponding to each pre-processed macro block to obtain a coded macro block corresponding to each pre-processed macro block.
5. An image processing method applied to an image receiving apparatus, the method comprising:
receiving target data of each macro block in claims 1-2 sent by an image sending device;
reducing the dimension of each macro block according to the gray level of the basic layer data of each macro block to obtain a frame image after dimension reduction;
amplifying the reduced frame image into a frame image with a target size;
and performing super-resolution reconstruction on the frame image with the target size and the at least one enhancement layer data to generate a target frame image.
6. The method of claim 5, wherein the super-resolution reconstruction of the target size frame image and the at least one enhancement layer data comprises:
and inputting the frame image with the target size and the at least one enhancement layer data into a pre-trained super-resolution reconstruction convolutional neural network (SRCNN) model.
7. The method according to claim 5, wherein each macroblock transmitted by the image transmission device is an encoded macroblock, and after receiving target data of each macroblock transmitted by the image transmission device, the method further comprises:
and decoding each macro block to generate a decoded macro block corresponding to each macro block.
8. An image processing apparatus, applied to an image transmission device, the apparatus comprising:
the frame image to be processed acquiring module is used for acquiring a frame image to be processed and dividing macro blocks of the frame image to be processed to generate at least one macro block;
a macroblock preprocessing module, configured to preprocess each macroblock of the at least one macroblock to obtain a preprocessed macroblock corresponding to each macroblock, where each preprocessed macroblock includes a DC component and at least one AC component;
the device comprises a preprocessing macro block layering module, a preprocessing macro block layering module and a control module, wherein the preprocessing macro block layering module is used for layering each preprocessing macro block according to a preset layering rule to obtain layering data corresponding to each preprocessing macro block, the layering data comprise a base layer and at least one enhancement layer, the DC component is the base layer, the AC component is divided into at least one enhancement layer according to importance degrees, and the importance degrees of the AC components in the same enhancement layer are the same;
the hierarchical data extraction module is used for extracting the hierarchical data corresponding to each preprocessed macro block according to a preset data extraction rule to obtain target data corresponding to each preprocessed macro block, wherein the target data comprises a base layer data and at least one enhancement layer data;
and the target data sending module is used for sending the target data corresponding to each preprocessed macro block to the image receiving equipment.
9. The apparatus of claim 8, wherein the macroblock preprocessing module is specifically configured to:
performing DCT (discrete cosine transformation) on each macro block in the at least one macro block to obtain a transformed macro block corresponding to each macro block, wherein the transformed macro block comprises a DC component and at least one AC component;
and quantizing each transformed macroblock to obtain a quantized macroblock corresponding to each transformed macroblock, wherein the quantized macroblock comprises a quantized DC component and at least one quantized AC component.
10. An image processing apparatus, applied to an image receiving device, the apparatus comprising:
a target data receiving module, configured to receive target data of each macro block according to claims 1-2, sent by an image sending device;
the macro block dimension reduction module is used for reducing the dimension of each macro block according to the gray level of the basic layer data of each macro block to obtain a frame image after dimension reduction;
the frame image amplifying module is used for amplifying the frame image subjected to the dimensionality reduction into a frame image with a target size;
and the target frame image generation module is used for performing super-resolution reconstruction on the frame image with the target size and the at least one enhancement layer data to generate a target frame image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010581960.7A CN111800633A (en) | 2020-06-23 | 2020-06-23 | Image processing method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010581960.7A CN111800633A (en) | 2020-06-23 | 2020-06-23 | Image processing method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111800633A true CN111800633A (en) | 2020-10-20 |
Family
ID=72804588
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010581960.7A Pending CN111800633A (en) | 2020-06-23 | 2020-06-23 | Image processing method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111800633A (en) |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2001251633A (en) * | 2001-03-12 | 2001-09-14 | Toshiba Corp | Moving picture coder and moving picture coding method |
| US20020122481A1 (en) * | 2001-01-24 | 2002-09-05 | Shinichi Mine | Transcoding method and transcoding apparatus |
| US20040264580A1 (en) * | 2003-03-17 | 2004-12-30 | Stmicroelectronics Asia Pacific Pte Ltd. | Decoder and method of decoding using pseudo two pass decoding and one pass encoding |
| JP3895764B1 (en) * | 2006-09-25 | 2007-03-22 | ５ｄｉｇｉｓｔａｒ株式会社 | Image processing system |
| CN101069433A (en) * | 2004-12-03 | 2007-11-07 | 三星电子株式会社 | Method and apparatus for encoding/decoding multi-layer video using DCT upsampling |
| CN101146224A (en) * | 2007-09-20 | 2008-03-19 | 合肥工业大学 | DCT coefficient hierarchical disorder method based on conversion domain in H.264 |
| US20090110054A1 (en) * | 2007-10-24 | 2009-04-30 | Samsung Electronics Co., Ltd. | Method, medium, and apparatus for encoding and/or decoding video |
| JP2012085256A (en) * | 2010-10-13 | 2012-04-26 | Samsung Electronics Co Ltd | Entropy encoding/decoding method and device for video processing of hierarchic structure and symbol encoding/decoding method and device therefor |
| US20140003505A1 (en) * | 2012-07-02 | 2014-01-02 | Nokia Corporation | Method and apparatus for video coding |
| US20150139304A1 (en) * | 2012-06-29 | 2015-05-21 | Sony Corporation | Image processing apparatus and method |
| CN106534859A (en) * | 2015-09-11 | 2017-03-22 | 中兴通讯股份有限公司 | Image transmission method and device based on SPICE protocol |
| US20170230685A1 (en) * | 2016-02-08 | 2017-08-10 | Canon Kabushiki Kaisha | Methods, devices, and computer programs for combining the use of intra-layer prediction and inter-layer prediction with scalability and screen content features |
| CN111277829A (en) * | 2020-02-25 | 2020-06-12 | 西安万像电子科技有限公司 | Encoding and decoding method and device |
-
2020
- 2020-06-23 CN CN202010581960.7A patent/CN111800633A/en active Pending
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020122481A1 (en) * | 2001-01-24 | 2002-09-05 | Shinichi Mine | Transcoding method and transcoding apparatus |
| JP2001251633A (en) * | 2001-03-12 | 2001-09-14 | Toshiba Corp | Moving picture coder and moving picture coding method |
| US20040264580A1 (en) * | 2003-03-17 | 2004-12-30 | Stmicroelectronics Asia Pacific Pte Ltd. | Decoder and method of decoding using pseudo two pass decoding and one pass encoding |
| CN101069433A (en) * | 2004-12-03 | 2007-11-07 | 三星电子株式会社 | Method and apparatus for encoding/decoding multi-layer video using DCT upsampling |
| JP3895764B1 (en) * | 2006-09-25 | 2007-03-22 | ５ｄｉｇｉｓｔａｒ株式会社 | Image processing system |
| CN101146224A (en) * | 2007-09-20 | 2008-03-19 | 合肥工业大学 | DCT coefficient hierarchical disorder method based on conversion domain in H.264 |
| US20090110054A1 (en) * | 2007-10-24 | 2009-04-30 | Samsung Electronics Co., Ltd. | Method, medium, and apparatus for encoding and/or decoding video |
| JP2012085256A (en) * | 2010-10-13 | 2012-04-26 | Samsung Electronics Co Ltd | Entropy encoding/decoding method and device for video processing of hierarchic structure and symbol encoding/decoding method and device therefor |
| US20150139304A1 (en) * | 2012-06-29 | 2015-05-21 | Sony Corporation | Image processing apparatus and method |
| US20140003505A1 (en) * | 2012-07-02 | 2014-01-02 | Nokia Corporation | Method and apparatus for video coding |
| CN106534859A (en) * | 2015-09-11 | 2017-03-22 | 中兴通讯股份有限公司 | Image transmission method and device based on SPICE protocol |
| US20170230685A1 (en) * | 2016-02-08 | 2017-08-10 | Canon Kabushiki Kaisha | Methods, devices, and computer programs for combining the use of intra-layer prediction and inter-layer prediction with scalability and screen content features |
| CN111277829A (en) * | 2020-02-25 | 2020-06-12 | 西安万像电子科技有限公司 | Encoding and decoding method and device |
Non-Patent Citations (3)
| Title |
|---|
| JASPREET BHATIA: "A novel image enhancement technique based on statistical analysis of DCT coefficients for JPEG compressed images", 《IEEE》, 8 September 2016 (2016-09-08) * |
| 张哲为: "基于感兴趣区域率失真优化的视频压缩编码通信系统设计", 《博士电子期刊》, 15 January 2020 (2020-01-15) * |
| 李永忠: "基于小波变换的空间可分级图像压缩编码算法", 计算机应用与软件, no. 02, 12 February 2005 (2005-02-12) * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110087092B (en) | Low-bit-rate video coding and decoding method based on image reconstruction convolutional neural network | |
| Li et al. | Learning convolutional networks for content-weighted image compression | |
| Zhou et al. | End-to-end Optimized Image Compression with Attention Mechanism. | |
| Cai et al. | Efficient variable rate image compression with multi-scale decomposition network | |
| Zhao et al. | Learning a virtual codec based on deep convolutional neural network to compress image | |
| CN113301347B (en) | HEVC high definition video coding optimization method | |
| CN113079378B (en) | Image processing method and device and electronic equipment | |
| CN111247797A (en) | Method and apparatus for image encoding and decoding | |
| EP4260558A1 (en) | Method and data processing system for lossy image or video encoding, transmission and decoding | |
| CN112333451A (en) | Intra-frame prediction method based on generation countermeasure network | |
| CN102917226B (en) | Intra-frame video coding method based on self-adaption downsampling and interpolation | |
| CN116916036A (en) | Video compression method, device and system | |
| Fu et al. | An extended hybrid image compression based on soft-to-hard quantification | |
| CN115604485A (en) | Video image decoding method and device | |
| US10015484B2 (en) | Adaptive scan device and method for scanning thereof | |
| CN115205117B (en) | Image reconstruction method and device, computer storage medium and electronic equipment | |
| CN116567240A (en) | Image compression method and system based on adaptive channel and spatial window entropy model | |
| CN111050170A (en) | Image compression system construction method, compression system and method based on GAN | |
| CN111800633A (en) | Image processing method and device | |
| CN115150628A (en) | Coarse-to-fine depth video coding method with super-prior guiding mode prediction | |
| CN105611288B (en) | A kind of low bit rate image sequence coding method based on Constrained interpolation technique | |
| US8306341B2 (en) | Image data compression apparatus and decoding apparatus | |
| Zhang et al. | Dual-layer image compression via adaptive downsampling and spatially varying upconversion | |
| CN114529442A (en) | Robust image watermarking method adopting two-stage precoding and wavelet network | |
| Kumar et al. | Lossy compression of color images using lifting scheme and prediction errors |
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 |