CN109102463A - A kind of super-resolution image reconstruction method and device - Google Patents
A kind of super-resolution image reconstruction method and device Download PDFInfo
- Publication number
- CN109102463A CN109102463A CN201810918370.1A CN201810918370A CN109102463A CN 109102463 A CN109102463 A CN 109102463A CN 201810918370 A CN201810918370 A CN 201810918370A CN 109102463 A CN109102463 A CN 109102463A
- Authority
- CN
- China
- Prior art keywords
- size
- layer
- convolutional neural
- picture element
- neural networks
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 238000012545 processing Methods 0.000 claims abstract description 55
- 238000013136 deep learning model Methods 0.000 claims abstract description 26
- 238000013527 convolutional neural network Methods 0.000 claims description 60
- 239000011159 matrix material Substances 0.000 claims description 36
- 238000013528 artificial neural network Methods 0.000 claims description 13
- 238000004590 computer program Methods 0.000 claims description 7
- 230000001537 neural effect Effects 0.000 claims description 7
- 238000004891 communication Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims 1
- 230000033001 locomotion Effects 0.000 abstract description 9
- 238000013461 design Methods 0.000 abstract description 6
- 238000010276 construction Methods 0.000 description 4
- 230000001052 transient effect Effects 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000004218 nerve net Anatomy 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4053—Scaling of whole images or parts thereof, e.g. expanding or contracting based on super-resolution, i.e. the output image resolution being higher than the sensor resolution
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The present invention provides a kind of super-resolution image reconstruction method and devices, comprising: the size of image to be detected after progress gray processing processing is set as first size, obtains first size image to be detected；First size image to be detected is input to first layer deep learning model, exports three picture element matrixs of first size；Three picture element matrixs are input to second layer deep learning model, export three picture element matrixs of the second size, the second size is greater than first size；Three picture element matrixs of the second size are integrated, super-resolution image is obtained.The processing speed that the present invention is converted into high-definition picture to low-resolution image is fast, without having to worry about object of which movement, it is high to rebuild degree, and design simply, processing speed is fast.
Description
Technical field
The present invention relates to image procossing and machine learning techniques field more particularly to a kind of super-resolution image reconstruction methods
And device.
Background technique
Requirement with people to picture pixels is higher and higher, and people increasingly praise highly the camera of high pixel, still,
It is limited the resolution ratio of the camera of high pixel, it needs the picture of low resolution being converted to high-resolution picture.
Currently, the method that Technique of Super-resolution Image Construction is mainly based upon frequency or airspace, but the side based on frequency domain
Method is only applicable in image the case where there is no local motions, only object mass motion, and only the noise in airspace is constant
In the case where be applicable in；And method design based on airspace is complicated, calculates that step is more, speed is slow.
Therefore, current Technique of Super-resolution Image Construction, which exists, is only applicable to object mass motion, and part may be not present
Movement or design are complicated, calculate step mostly with slow-footed problem.
Summary of the invention
In order to solve current Technique of Super-resolution Image Construction in the presence of object mass motion is only applicable to, part may be not present
Movement or design are complicated, calculate step mostly with slow-footed problem, on the one hand, the present invention provides a kind of super-resolution image weights
Construction method, comprising:
The size of image to be detected after progress gray processing processing is set as first size, it is to be detected to obtain first size
Image；
First size image to be detected is input to first layer deep learning model, exports three pixel squares of first size
Battle array；
Three picture element matrixs are input to second layer deep learning model, export three picture element matrixs of the second size, the second ruler
It is very little to be greater than first size；
Three picture element matrixs of the second size are integrated, super-resolution image is obtained.
Preferably, the size of image to be detected after progress gray processing processing is set as first size, obtains the first ruler
Very little image to be detected, comprising:
Image to be detected is subjected to gray processing processing, image to be detected is made to become the channel R, the channel G and B from RGB triple channel
Channel obtains image to be detected after gray processing processing；
The dimension enlargement of image to be detected after being handled gray processing using bilinear interpolation is contracted to the first ruler
It is very little.
Preferably, three picture element matrixs include the channel R picture element matrix, the channel G picture element matrix and channel B picture element matrix.
Preferably, first layer deep learning model includes the first convolutional neural networks arranged side by side, the second convolutional neural networks
With third convolutional neural networks, the first convolutional neural networks, the second convolutional neural networks and third convolutional neural networks are distinguished
For 3 layers of convolutional neural networks；First convolutional neural networks export the channel the R picture element matrix of first size, the second convolution nerve net
Network exports the channel the G picture element matrix of first size, and third convolutional neural networks export the channel B picture element matrix of first size.
Preferably, 3 layers of convolutional neural networks include the first layer convolutional layer, second layer convolutional layer and third layer successively carried out
Convolutional layer；The size of the convolution kernel of first layer convolutional layer is 5*5, and the size of the convolution kernel of second layer convolutional layer is 4*4, third layer
The size of the convolution kernel of convolutional layer is 3*3, the number of the convolution kernel of first layer convolutional layer, second layer convolutional layer and third layer convolutional layer
Amount is 64.
Preferably, second layer deep learning network includes Volume Four product neural network, the 5th convolutional neural networks arranged side by side
With the 6th convolutional neural networks；Volume Four is accumulated neural network, the 5th convolutional neural networks and the 6th convolutional neural networks and is distinguished
For 18 layers of convolutional neural networks；Volume Four product neural network exports the channel the R picture element matrix of the second size, the 5th convolutional Neural net
Network exports the channel the G picture element matrix of the second size, and the 6th convolutional neural networks export the channel B picture element matrix of the second size.
Preferably, the size of the convolution kernel of each layer of convolutional layer of 18 layers of convolutional neural networks is 3*3, the number of convolution kernel
Amount is 32；18 layers of convolutional neural networks are the convolutional neural networks of the structure containing residual error, the convolutional Neural net of the structure containing residual error
The residual error of network connects the input from 18 layers of convolutional neural networks up to output.
On the other hand, the present invention also provides a kind of super-resolution image reconstruction devices, comprising:
Module is obtained, for obtaining image to be detected；
First processing module is set as first size, by the first ruler after image to be detected is carried out gray processing processing
Very little image to be detected is input to first layer deep learning model, exports three picture element matrixs of first size；
Three picture element matrixs are input to second layer deep learning model, export three pictures of the second size by Second processing module
Prime matrix, the second size are greater than first size；
Module is integrated, three picture element matrixs of the second size are integrated, obtains super-resolution image.
Another aspect, a kind of electronic equipment for super-resolution image reconstruction, comprising:
Memory and processor, processor and memory complete mutual communication by bus；Memory is stored with can
The program instruction being executed by processor, the instruction of processor caller are able to carry out above-mentioned method method.
In another aspect, being stored thereon with computer program the present invention provides a kind of computer readable storage medium, calculate
Machine program realizes above-mentioned method when being executed by processor.
The present invention provides a kind of super-resolution image reconstruction method and devices, will be grey by first layer deep learning model
After degreeization processing and it is set as image to be detected of first size, handles three picture element matrixs to keep first size；Pass through again
Three picture element matrixs of first size are handled three picture element matrixs for the second size by second layer depth model, so that three pixel squares
Battle array is apparent, finally integrates three picture element matrixs of the second size, obtains super-resolution image.The present invention is to low-resolution image
The processing speed for being converted into high-definition picture is fast, and without having to worry about object of which movement, it is high to rebuild degree, and designs simple, processing speed
Degree is fast.
Detailed description of the invention
Fig. 1 is to be illustrated according to a kind of process of super-resolution image reconstruction method of a preferred embodiment of the present invention
Figure；
Fig. 2 is the structural representation according to a kind of super-resolution image reconstruction device of a preferred embodiment of the present invention
Figure；
Fig. 3 is a kind of electronic equipment for super-resolution image reconstruction according to a preferred embodiment of the present invention
Structural schematic diagram.
Specific embodiment
With reference to the accompanying drawings and examples, specific embodiments of the present invention will be described in further detail.Implement below
Example is not intended to limit the scope of the invention for illustrating the present invention.
Fig. 1 is to be illustrated according to a kind of process of super-resolution image reconstruction method of a preferred embodiment of the present invention
Figure, as shown in Figure 1, the embodiment of the invention provides a kind of super-resolution image reconstruction methods, comprising:
Step S101, the size of image to be detected after progress gray processing processing is set as first size, obtains first
Size image to be detected；
Step S102, first size image to be detected is input to first layer deep learning model, exports three picture element matrixs,
The size of three picture element matrixs keeps first size；
Step S103, three picture element matrixs are input to second layer deep learning model, export three pixel squares of the second size
Battle array, the second size are greater than first size；
Step S104, three picture element matrixs of the second size are integrated, obtains super-resolution image.
Specifically, it is necessary first to obtain image to be detected, point that image to be detected is usually shot by camera or video camera
The lower image of resolution, if desired sees the details of image to be detected clearly, needs to carry out super-resolution image reconstruction to it.
Further, in order to enable image to be detected to facilitate processing, image to be detected is first subjected to gray processing processing, by three
Channel becomes single channel, and the size of the image handled through gray processing is uniformly set as first size image to be detected, obtains the
One size image to be detected；Then first size image to be detected is input to first layer deep learning model, output three is not
Same picture element matrix, i.e. three picture element matrixs, and the size of three picture element matrixs still maintains first size；Next, again by three pictures
Prime matrix is input to second layer deep learning model, so that the size of three picture element matrixs is become the second size from first size, substantially
The size of three picture element matrixs is improved, finally integrates three picture element matrixs of the second size, obtains super-resolution image.
Wherein, first size is less than 1920*1080, and the second size is not less than 1920*1080；For example, first size is
640*480, second having a size of 1080*1920.
It should be noted that three picture element matrixs include the channel R picture element matrix, the channel G picture element matrix and channel B pixel square
Battle array.
Based on the above embodiment, first layer deep learning model includes the first convolutional neural networks arranged side by side, the second convolution
Neural network and third convolutional neural networks, the first convolutional neural networks, the second convolutional neural networks and third convolutional Neural net
Network is respectively 3 layers of convolutional neural networks；First convolutional neural networks export the channel the R picture element matrix of first size, volume Two
The channel the G picture element matrix of product neural network output first size, third convolutional neural networks export the channel B pixel of first size
Matrix.
Further, 3 layers of convolutional neural networks include the first layer convolutional layer, second layer convolutional layer and third successively carried out
Layer convolutional layer；The size of the convolution kernel of first layer convolutional layer is 5*5, and the size of the convolution kernel of second layer convolutional layer is 4*4, third
The size of the convolution kernel of layer convolutional layer is 3*3, the convolution kernel of first layer convolutional layer, second layer convolutional layer and third layer convolutional layer
Quantity is 64.
Based on the above embodiment, second layer deep learning network includes Volume Four product neural network, the 5th convolution arranged side by side
Neural network and the 6th convolutional neural networks；Volume Four accumulates neural network, the 5th convolutional neural networks and the 6th convolutional Neural net
Network is respectively 18 layers of convolutional neural networks；The channel the R picture element matrix of Volume Four product neural network the second size of output, volume five
Product neural network exports the channel the G picture element matrix of the second size, and the 6th convolutional neural networks export the channel B pixel of the second size
Matrix.
Further, the input of Volume Four product neural network is the channel the R picture element matrix of first size, the 5th convolutional Neural
The input of network is the channel the G picture element matrix of first size, and the input of the 6th convolutional neural networks is the channel B picture of first size
Prime matrix.
Further, the size of the convolution kernel of each layer of convolutional layer of 18 layers of convolutional neural networks is 3*3, convolution kernel
Quantity is 32；18 layers of convolutional neural networks are the convolutional neural networks of the structure containing residual error, the convolutional Neural of the structure containing residual error
The residual error of network connects the input from 18 layers of convolutional neural networks up to output.
Fig. 2 is the structural representation according to a kind of super-resolution image reconstruction device of a preferred embodiment of the present invention
Figure, as shown in Fig. 2, the device includes obtaining module the embodiment of the invention provides a kind of super-resolution image reconstruction device
201, first processing module 202, Second processing module 203 and module 204 is integrated, in which:
Module 201 is obtained, for obtaining image to be detected；
First processing module 202 is set as first size after image to be detected is carried out gray processing processing, by the
One size image to be detected is input to first layer deep learning model, exports three picture element matrixs, and the size of three picture element matrixs is kept
First size；
Three picture element matrixs are input to second layer deep learning model, export the three of the second size by Second processing module 203
Picture element matrix, the second size are greater than first size；
Module 204 is integrated, three picture element matrixs of the second size are integrated, obtains super-resolution image.
Fig. 3 is a kind of electronic equipment for super-resolution image reconstruction according to a preferred embodiment of the present invention
Structural schematic diagram, as shown in figure 3, the present invention provides a kind of electronic equipment for super-resolution image reconstruction, the equipment packet
Include processor 301, memory 302 and bus 303；
Wherein, processor 301 and memory 302 complete mutual communication by the bus 303；
Processor 301 is used to call the program instruction in memory 302, to execute provided by above-mentioned each method embodiment
Method, for example,
The size of image to be detected after progress gray processing processing is set as first size, it is to be detected to obtain first size
Image；
First size image to be detected is input to first layer deep learning model, exports three pixel squares of first size
Battle array；
Three picture element matrixs are input to second layer deep learning model, export three picture element matrixs of the second size, the second ruler
It is very little to be greater than first size；
Three picture element matrixs of the second size are integrated, super-resolution image is obtained.
The embodiment of the present invention discloses a kind of computer program product, and the computer program product is non-transient including being stored in
Computer program on computer readable storage medium, the computer program include program instruction, when described program instructs quilt
When computer executes, computer is able to carry out method provided by above-mentioned each method embodiment, for example,
The size of image to be detected after progress gray processing processing is set as first size, it is to be detected to obtain first size
Image；
First size image to be detected is input to first layer deep learning model, exports three picture element matrixs, three pixel squares
The size of battle array keeps first size；
Three picture element matrixs are input to second layer deep learning model, export three picture element matrixs of the second size, the second ruler
It is very little to be greater than first size；
Three picture element matrixs of the second size are integrated, super-resolution image is obtained.
The embodiment of the present invention provides a kind of non-transient computer readable storage medium, the non-transient computer readable storage
Medium storing computer instruction, the computer instruction make the computer execute side provided by above-mentioned each method embodiment
Method, for example,
The size of image to be detected after progress gray processing processing is set as first size, it is to be detected to obtain first size
Image；
First size image to be detected is input to first layer deep learning model, exports three picture element matrixs, three pixel squares
The size of battle array keeps first size；
Three picture element matrixs are input to second layer deep learning model, export three picture element matrixs of the second size, the second ruler
It is very little to be greater than first size；
Three picture element matrixs of the second size are integrated, super-resolution image is obtained.
Those of ordinary skill in the art will appreciate that: realize that all or part of the steps of above method embodiment can pass through
The relevant hardware of program instruction is completed, and program above-mentioned can be stored in a computer readable storage medium, the program
When being executed, step including the steps of the foregoing method embodiments is executed；And storage medium above-mentioned includes: ROM, RAM, magnetic disk or light
The various media that can store program code such as disk.
The embodiments such as device and equipment described above are only schematical, wherein described be used as separate part description
Unit may or may not be physically separated, component shown as a unit may or may not be
Physical unit, it can it is in one place, or may be distributed over multiple network units.It can be according to the actual needs
Some or all of the modules therein is selected to realize the purpose of the embodiment of the present invention.Those of ordinary skill in the art are not
In the case where paying creative labor, it can understand and implement.
Through the above description of the embodiments, those skilled in the art can be understood that each embodiment can
It realizes by means of software and necessary general hardware platform, naturally it is also possible to pass through hardware.Based on this understanding, on
Stating technical solution, substantially the part that contributes to existing technology can be embodied in the form of software products in other words, should
Computer software product may be stored in a computer readable storage medium, such as ROM/RAM, magnetic disk, CD, including several fingers
It enables and using so that a computer equipment (can be personal computer, server or the network equipment etc.) executes each implementation
Method described in certain parts of example or embodiment.
The present invention provides a kind of super-resolution image reconstruction method and devices, will be grey by first layer deep learning model
After degreeization processing and it is set as image to be detected of first size, handles three picture element matrixs to keep first size；Pass through again
Three picture element matrixs of first size are handled three picture element matrixs for the second size by second layer depth model, so that three pixel squares
Battle array is apparent, finally integrates three picture element matrixs of the second size, obtains super-resolution image.The present invention is to low-resolution image
The processing speed for being converted into high-definition picture is fast, and without having to worry about object of which movement, it is high to rebuild degree, and designs simple, processing speed
Degree is fast.
Finally, method of the invention is only preferable embodiment, it is not intended to limit the scope of the present invention.It is all
Within the spirit and principles in the present invention, any modification, equivalent replacement, improvement and so on should be included in protection of the invention
Within the scope of.
Claims (10)
1. a kind of super-resolution image reconstruction method characterized by comprising
The size of image to be detected after progress gray processing processing is set as first size, obtains first size mapping to be checked
Picture；
Described first size image to be detected is input to first layer deep learning model, exports three pixel squares of first size
Battle array；
Three picture element matrix is input to second layer deep learning model, exports three picture element matrixs of the second size, described the
Two sizes are greater than the first size；
Three picture element matrixs of second size are integrated, super-resolution image is obtained.
2. a kind of super-resolution image reconstruction method according to claim 1, which is characterized in that described to carry out gray processing
The size of image to be detected after processing is set as first size, obtains first size image to be detected, comprising:
Described image to be detected is subjected to gray processing processing, described image to be detected is made to become the channel R, the channel G from RGB triple channel
And channel B, obtain image to be detected after the gray processing processing；
By the dimension enlargement of image to be detected after gray processing processing or described the is contracted to using bilinear interpolation
One size.
3. a kind of super-resolution image reconstruction method according to claim 2, which is characterized in that the three picture element matrixs packet
Include the channel R picture element matrix, the channel G picture element matrix and channel B picture element matrix.
4. a kind of super-resolution image reconstruction method according to claim 3, which is characterized in that first layer depth
Practising model includes the first convolutional neural networks arranged side by side, the second convolutional neural networks and third convolutional neural networks, and described first
Convolutional neural networks, the second convolutional neural networks and third convolutional neural networks are respectively 3 layers of convolutional neural networks；Described
One convolutional neural networks export the channel the R picture element matrix of first size, the G of the second convolutional neural networks output first size
Channel picture element matrix, the channel B picture element matrix of the third convolutional neural networks output first size.
5. a kind of super-resolution image reconstruction method according to claim 4, which is characterized in that 3 layers of convolutional Neural
Network includes the first layer convolutional layer, second layer convolutional layer and third layer convolutional layer successively carried out；The first layer convolutional layer
The size of convolution kernel is 5*5, and the size of the convolution kernel of the second layer convolutional layer is 4*4, the convolution of the third layer convolutional layer
The size of core is 3*3, and the quantity of the convolution kernel of the first layer convolutional layer, second layer convolutional layer and third layer convolutional layer is 64
It is a.
6. a kind of super-resolution image reconstruction method according to claim 3, which is characterized in that second layer depth
Practising network includes Volume Four product neural network, the 5th convolutional neural networks and the 6th convolutional neural networks arranged side by side；Described 4th
Convolutional neural networks, the 5th convolutional neural networks and the 6th convolutional neural networks are respectively 18 layers of convolutional neural networks；It is described
Volume Four product neural network exports the channel the R picture element matrix of the second size, and the 5th convolutional neural networks export the second size
The channel G picture element matrix, the 6th convolutional neural networks export the channel B picture element matrix of the second size.
7. a kind of super-resolution image reconstruction method according to claim 6, which is characterized in that 18 layers of convolutional Neural
The size of the convolution kernel of each layer of convolutional layer of network is 3*3, and the quantity of convolution kernel is 32；18 layers of convolutional Neural
Network is the convolutional neural networks of the structure containing residual error, and the residual error of the convolutional neural networks of the structure containing residual error is connected from described 18
The input of layer convolutional neural networks is until output.
8. a kind of super-resolution image reconstruction device characterized by comprising
Module is obtained, for obtaining image to be detected；
First processing module is set as first size, by the first ruler after described image to be detected is carried out gray processing processing
Very little image to be detected is input to first layer deep learning model, exports three picture element matrixs of first size；
Three picture element matrix is input to second layer deep learning model, exports three pictures of the second size by Second processing module
Prime matrix, second size are greater than the first size；
Module is integrated, three picture element matrixs of second size are integrated, obtains super-resolution image.
9. a kind of electronic equipment for super-resolution image reconstruction characterized by comprising
Memory and processor, the processor and the memory complete mutual communication by bus；The memory
It is stored with the program instruction that can be executed by the processor, the processor calls described program instruction to be able to carry out right such as and wants
Seek 1 to 7 any method.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program
The method as described in claim 1 to 7 is any is realized when being executed by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810918370.1A CN109102463B (en) | 2018-08-13 | 2018-08-13 | Super-resolution image reconstruction method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810918370.1A CN109102463B (en) | 2018-08-13 | 2018-08-13 | Super-resolution image reconstruction method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109102463A true CN109102463A (en) | 2018-12-28 |
| CN109102463B CN109102463B (en) | 2023-01-24 |
Family
ID=64849380
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810918370.1A Active CN109102463B (en) | 2018-08-13 | 2018-08-13 | Super-resolution image reconstruction method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109102463B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021047118A1 (en) * | 2019-09-12 | 2021-03-18 | 浪潮电子信息产业股份有限公司 | Image processing method, device and system |
| CN114066722A (en) * | 2021-11-03 | 2022-02-18 | 北京字节跳动网络技术有限公司 | Method and device for acquiring image and electronic equipment |
| CN114549328A (en) * | 2022-04-24 | 2022-05-27 | 西南财经大学 | JPG image super-resolution recovery method, computer-readable storage medium and terminal |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2014033415A (en) * | 2012-08-06 | 2014-02-20 | Canon Inc | Image processor, image processing method and imaging apparatus |
| CN105046672A (en) * | 2015-06-30 | 2015-11-11 | 北京工业大学 | Method for image super-resolution reconstruction |
| CN107358575A (en) * | 2017-06-08 | 2017-11-17 | 清华大学 | A kind of single image super resolution ratio reconstruction method based on depth residual error network |
| CN108090870A (en) * | 2017-12-13 | 2018-05-29 | 苏州长风航空电子有限公司 | A kind of infrared image super resolution ratio reconstruction method based on thaumatropy self similarity |
| CN108259997A (en) * | 2018-04-02 | 2018-07-06 | 腾讯科技（深圳）有限公司 | Image correlation process method and device, intelligent terminal, server, storage medium |
-
2018
- 2018-08-13 CN CN201810918370.1A patent/CN109102463B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2014033415A (en) * | 2012-08-06 | 2014-02-20 | Canon Inc | Image processor, image processing method and imaging apparatus |
| CN105046672A (en) * | 2015-06-30 | 2015-11-11 | 北京工业大学 | Method for image super-resolution reconstruction |
| CN107358575A (en) * | 2017-06-08 | 2017-11-17 | 清华大学 | A kind of single image super resolution ratio reconstruction method based on depth residual error network |
| CN108090870A (en) * | 2017-12-13 | 2018-05-29 | 苏州长风航空电子有限公司 | A kind of infrared image super resolution ratio reconstruction method based on thaumatropy self similarity |
| CN108259997A (en) * | 2018-04-02 | 2018-07-06 | 腾讯科技（深圳）有限公司 | Image correlation process method and device, intelligent terminal, server, storage medium |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021047118A1 (en) * | 2019-09-12 | 2021-03-18 | 浪潮电子信息产业股份有限公司 | Image processing method, device and system |
| US11614964B2 (en) | 2019-09-12 | 2023-03-28 | Inspur Electronic Information Industry Co., Ltd. | Deep-learning-based image processing method and system |
| CN114066722A (en) * | 2021-11-03 | 2022-02-18 | 北京字节跳动网络技术有限公司 | Method and device for acquiring image and electronic equipment |
| CN114066722B (en) * | 2021-11-03 | 2024-03-19 | 抖音视界有限公司 | Method and device for acquiring image and electronic equipment |
| CN114549328A (en) * | 2022-04-24 | 2022-05-27 | 西南财经大学 | JPG image super-resolution recovery method, computer-readable storage medium and terminal |
| CN114549328B (en) * | 2022-04-24 | 2022-07-22 | 西南财经大学 | JPG image super-resolution restoration method, computer readable storage medium and terminal |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109102463B (en) | 2023-01-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11354785B2 (en) | Image processing method and device, storage medium and electronic device | |
| Zhao et al. | Efficient image super-resolution using pixel attention | |
| US20200372609A1 (en) | Super-resolution video reconstruction method, device, apparatus and computer-readable storage medium | |
| CN108022212B (en) | High-resolution picture generation method, generation device and storage medium | |
| CN109146788A (en) | Super-resolution image reconstruction method and device based on deep learning | |
| CN111080528A (en) | Image super-resolution and model training method, device, electronic equipment and medium | |
| CN111194458A (en) | Image signal processor for processing image | |
| CN109102463A (en) | A kind of super-resolution image reconstruction method and device | |
| JP2019501454A (en) | Computer system and method for upsampling images | |
| CN107977414A (en) | Image Style Transfer method and its system based on deep learning | |
| CN112801901A (en) | Image deblurring algorithm based on block multi-scale convolution neural network | |
| CN111242846A (en) | Fine-grained scale image super-resolution method based on non-local enhancement network | |
| CN109117944B (en) | Super-resolution reconstruction method and system for ship target remote sensing image | |
| CN109993712A (en) | Training method, image processing method and the relevant device of image processing model | |
| CN110062282A (en) | A kind of super-resolution video method for reconstructing, device and electronic equipment | |
| CN109871841B (en) | Image processing method, device, terminal and storage medium | |
| JP7352748B2 (en) | Three-dimensional reconstruction method, device, equipment and storage medium | |
| CN111510739B (en) | Video transmission method and device | |
| CN110428382A (en) | A kind of efficient video Enhancement Method, device and storage medium for mobile terminal | |
| CN100474775C (en) | Finite impulse response filter method and apparatus | |
| CN110166798B (en) | Down-conversion method and device based on 4K HDR editing | |
| CN105427235B (en) | A kind of image browsing method and system | |
| CN105224538B (en) | The dithering process method and apparatus of image | |
| WO2023246403A1 (en) | Model training method, watermark restoration method, and related device | |
| US20230060988A1 (en) | Image processing device and method |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20230111 Address after: 215000 unit 2-b702, creative industry park, 328 Xinghu street, Suzhou Industrial Park, Jiangsu Province Applicant after: SUZHOU FEISOU TECHNOLOGY Co.,Ltd. Address before: Room 1216, 12 / F, Beijing Beiyou science and technology and cultural exchange center, 10 Xitucheng Road, Haidian District, Beijing, 100876 Applicant before: BEIJING FEISOU TECHNOLOGY CO.,LTD. |
|
| TA01 | Transfer of patent application right |