CN117218040A - Method and system for removing reflection of multiple images based on space-time information - Google Patents
Method and system for removing reflection of multiple images based on space-time information Download PDFInfo
- Publication number
- CN117218040A CN117218040A CN202311434745.4A CN202311434745A CN117218040A CN 117218040 A CN117218040 A CN 117218040A CN 202311434745 A CN202311434745 A CN 202311434745A CN 117218040 A CN117218040 A CN 117218040A
- Authority
- CN
- China
- Prior art keywords
- image
- images
- transmission
- toushe
- zhang
- 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
- 238000000034 method Methods 0.000 title claims abstract description 39
- 230000005540 biological transmission Effects 0.000 claims abstract description 112
- 238000005457 optimization Methods 0.000 claims abstract description 25
- 238000001028 reflection method Methods 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 5
- 238000003860 storage Methods 0.000 claims description 5
- 230000006870 function Effects 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 10
- 125000004122 cyclic group Chemical group 0.000 abstract description 3
- 238000012549 training Methods 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 6
- 230000009466 transformation Effects 0.000 description 4
- 238000013135 deep learning Methods 0.000 description 3
- 238000006073 displacement reaction Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 230000005489 elastic deformation Effects 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 238000013178 mathematical model Methods 0.000 description 2
- 238000013519 translation Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000007667 floating Methods 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
Abstract
The invention discloses a method and a system for removing reflection of multiple images based on space-time information, wherein the method comprises the following steps: acquiring an image sequence comprising a plurality of images, extracting front and back images in the sequence as input images, and initializing a transmission image of the input images; calculating deformation between two transmission images by using a para-LAP algorithm to obtain motion information and brightness relation between the two images, and aligning the two images; fixing the front transmission image, calculating the optimal solution of the rear Zhang Toushe image according to an optimization equation, and updating; based on the updated post Zhang Toushe image, calculating an optimal solution of the front Zhang Toushe image according to an optimization equation and updating; and based on the updated two transmission images, continuously performing the alignment and updating operation in a cyclic iteration mode until the maximum cyclic iteration times are reached, and outputting the finally updated transmission images. According to the invention, the antireflection of a plurality of images with better effect is realized according to the motion information between the front image and the rear image and the space information of the single image.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a method and a system for removing reflection of multiple images based on space-time information.
Background
With the continuous progress of image processing technology, people have increasingly higher requirements on the quality of images shot by imaging devices such as mobile phones and tablet computers. However, if an object made of reflective material such as glass exists in the shooting scene, the reflective image will be superimposed on the image shot by the imaging device, which further results in a decrease in the definition of the transmission image, resulting in poor quality of the shot image. Therefore, removing the reflection of light in an image is one of the keys to improve the image quality.
Existing image retroreflective techniques are generally classified into single image retroreflective techniques and multiple image retroreflective techniques. The single-image antireflection technology is realized based on a mathematical model of i=t+r+n, where I is an image containing reflection, T is a transmission image, i.e. an image without reflection, R is a reflection image, and N is noise. The task of image anti-reflection is to recover the non-reflection image T from the input image I. For the single image retroreflective task, T, R and N are both unknown, only I is known, the number of unknown parameters is greater than the number of known parameters, and therefore the single image retroreflective problem is an ill-posed problem. Existing single image anti-reflection techniques use the a priori information of I, T, R and add constraints on I, T, R to solve this uncertainty problem. However, these prior information cannot describe the inherent features of the image in all aspects, so that the conventional single image anti-reflection technology cannot effectively solve the anti-reflection problem, and the effect is poor.
The existing multi-image anti-reflection technology needs to shoot two or more images with different angles, and uses motion information or equipment information between the images to remove the reflection. Compared with the single-image antireflection technology, the method has the advantages that more information is available in the process of antireflection of a plurality of images, and a better antireflection effect is usually achieved, the currently commonly adopted multi-image antireflection algorithm is to decompose images into a projection layer and a reflection layer by utilizing motion information among a series of images, but the method is not applicable to videos or image sequences containing moving objects.
In recent years, with the rapid development of deep learning, more and more image antireflection technologies based on deep learning are proposed, and a better antireflection effect is obtained. However, such techniques require a significant amount of additional training data and training time, and the training process relies on high performance computing devices, limiting their use in real life and production.
Disclosure of Invention
In order to solve the defects in the prior art, the invention provides a multi-image antireflection method and a system based on space-time information, which adopt an accurate and rapid image registration method-a parameterized all-pass filter (parametric local all-pass, para-LAP) method to align transmission layers of two images and carry out sparse constraint on gradients of reflection layers between the two images so as to ensure the spatial smoothness of an antireflection image, thereby completing the antireflection task of the multi-image and obtaining the antireflection image with better effect.
In a first aspect, the present invention provides a method for the retroreflective of multiple images based on temporal and spatial information.
A method for removing light from a plurality of images based on spatio-temporal information, comprising:
acquiring an image sequence comprising a plurality of images, extracting a front image and a rear image in the image sequence, taking the extracted two images as input images, and initializing a transmission image of the input images;
calculating deformation between the two transmission images by using a para-LAP algorithm based on the transmission images initialized by the two input images to obtain motion information and brightness relation between the two transmission images, and further aligning the two transmission images;
fixing the front transmission image, calculating the optimal solution of the rear Zhang Toushe image according to an optimization equation, and updating the rear Zhang Toushe image; based on the updated post Zhang Toushe image, calculating an optimal solution of the pre-Zhang Toushe image according to an optimization equation, and updating the pre-transmission image;
based on the updated two transmission images, the alignment and updating operation is continuously and circularly executed until the maximum number of circulation iteration times is reached, and finally updated transmission images are output, namely, a plurality of final anti-reflection images are output.
In a second aspect, the present invention provides a multi-image retroreflective system based on spatio-temporal information.
A spatio-temporal information based multi-image retroreflective system comprising:
the image acquisition and preprocessing module is used for acquiring an image sequence comprising a plurality of images, extracting front and rear images in the image sequence, taking the extracted two images as input images, and initializing a transmission image of the input images;
the image alignment module is used for calculating deformation between the two transmission images based on the transmission images initialized by the two input images by using a para-LAP algorithm to obtain motion information and brightness relation between the two transmission images, so as to align the two transmission images;
the image processing module is used for fixing the previous transmission image, calculating the optimal solution of the post Zhang Toushe image according to the optimization equation, and updating the post Zhang Toushe image; based on the updated post Zhang Toushe image, calculating an optimal solution of the pre-Zhang Toushe image according to an optimization equation, and updating the pre-transmission image;
and the anti-reflection image output module is used for continuously and circularly executing the alignment and updating operation based on the two updated transmission images until the maximum number of the circular iterations is reached, and outputting the finally updated transmission images, namely outputting a plurality of final anti-reflection images.
In a third aspect, the present disclosure also provides an electronic device comprising a memory and a processor, and computer instructions stored on the memory and running on the processor, which when executed by the processor, perform the steps of the method of the first aspect.
In a fourth aspect, the present disclosure also provides a computer readable storage medium storing computer instructions which, when executed by a processor, perform the steps of the method of the first aspect.
The one or more of the above technical solutions have the following beneficial effects:
1. the invention provides a method and a system for removing reflection of multiple images based on space-time information, which adopt an accurate and rapid image registration method-a parameterized all-pass filtering method according to motion information between a front image and a rear image and space information of a single image, align transmission layers of the two images, sparsely constrain gradients of reflection layers between the two images, ensure the spatial smoothness of the removed reflection images, and complete the task of removing reflection of the multiple images by continuous loop iteration solution to obtain multiple removed reflection images with better effects; compared with the traditional single image deblurring algorithm, the method utilizes more information in a plurality of images, such as time and space information, and has better antireflection effect.
2. According to the invention, by combining the para-LAP algorithm, even if brightness difference exists between two input images, the motion information between the images can be accurately estimated; and the method is not dependent on training data and high-performance computing equipment, does not need extra training time, and is economical and efficient.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention.
Fig. 1 is a flowchart of a method for removing light from a plurality of images based on spatio-temporal information according to an embodiment of the present invention.
Detailed Description
It should be noted that the following detailed description is exemplary only for the purpose of describing particular embodiments and is intended to provide further explanation of the invention and is not intended to limit exemplary embodiments according to the invention. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Furthermore, it will be further understood that the terms "comprises" and/or "comprising," when used in this specification, specify the presence of stated features, steps, operations, devices, components, and/or groups thereof.
Example 1
Aiming at the problems that the single image anti-reflection method in the existing anti-image anti-reflection method is poor in anti-reflection effect, the multi-image anti-reflection method is not applicable to video and image sequences including moving objects, the anti-reflection method based on deep learning is high in limiting conditions and strong in dependence, and the like, the embodiment provides the multi-image anti-reflection method based on space-time information, as shown in fig. 1, and the method comprises the following steps:
step S1, acquiring an image sequence comprising a plurality of images, extracting front and rear images in the image sequence, taking the extracted two images as input images, and initializing a transmission image of the input images;
s2, calculating deformation between two transmission images based on the transmission images initialized by the two input images by using a para-LAP algorithm to obtain motion information and brightness relation between the two transmission images, and further aligning the two transmission images;
s3, fixing the front transmission image, calculating an optimal solution of the rear Zhang Toushe image according to an optimization equation, and updating the rear Zhang Toushe image; based on the updated post Zhang Toushe image, calculating an optimal solution of the pre-Zhang Toushe image according to an optimization equation, and updating the pre-transmission image;
and step S4, based on the two updated transmission images, performing alignment and updating operations continuously and circularly until the maximum number of circulation iterations is reached, and outputting a finally updated transmission image, namely outputting a plurality of final anti-reflection images.
The method for removing light from a plurality of images based on spatio-temporal information according to this embodiment will be described in more detail below.
In step S1, an image sequence including a plurality of images is acquired, two front and rear images in the image sequence are extracted, the extracted two images are used as input images, and a transmission image of the input images is initialized. For a single image, the mathematical model can be used to: i=t+r+n, where I means an image containing reflection, i.e., an input image, T means a transmission image, i.e., an image after reflection removal, R means a reflection image, and N means random noise. The present embodiment aims at extracting the reflected light image T from the input image I.
In the present embodiment, an image sequence including a plurality of images is obtained, the images in the image sequence are images including reflection, and the front and rear images in the image sequence are extracted and recorded as an input image I 1 And I 2 . Initializing an input image I 1 And I 2 Is respectively marked as a transmission image T 1 And T 2 Initialization here refers to the direct assignment of the input image to the transmission image, i.e. the input image I 1 As an initialized transmission image T 1 Input image I 2 As an initialized transmission image T 2 . At the same time, initialize the parameter lambda s And lambda (lambda) t Setting the maximum iteration number beta max 。
In step S2, based on the transmission images initialized by the two input images, deformation between the two transmission images is calculated by using a para-LAP algorithm, so that motion information and brightness relation between the two transmission images are obtained, and the two transmission images are aligned.
Specifically, for the initialized transmission image T 1 And T 2 Calculating deformation between two transmission images by using para-LAP algorithm to obtain motion information (u) between the two transmission images x (x,y),u y (x, y)) and a luminance relationship F, wherein the motion information refers to displacement information between each corresponding pixel point in the two transmission images, u x (x, y) represents the abscissa displacement for pixel point (x, y), u y (x, y) represents the ordinate displacement for the pixel point (x, y).
The para-LAP algorithm is an image registration algorithm, and deformation between two images is acquired through image registration. The para-LAP algorithm is as follows: when no gray level difference exists between the reference image and the floating image, local transformation in smooth space transformation can be approximated by translation transformation, the translation transformation is equivalent to full-pass filtering of an image block, a local full-pass filter is further obtained through estimation, then local deformation is extracted from the partial full-pass filter, and an elastic deformation field is obtained through window sliding; and finally, fitting the elastic deformation field by using a linear equation containing few parameters, thereby obtaining a smooth deformation field.
Further, according to the calculated motion information, the two transmission images are aligned, i.e. the transmission image T 2 To the transmission image T 1 Alignment. By the accurate and rapid image registration method, the projection layers of the two images are aligned, and the para-LAP algorithm is adopted for calculation, so that the problem of image brightness is considered, and even if the brightness of the two images is different, the two images can be accurately aligned.
In step S3, fixing the previous transmission image, calculating the optimal solution of the subsequent Zhang Toushe image according to an optimization equation, and updating the subsequent Zhang Toushe image; based on the updated post Zhang Toushe image, an optimal solution for the pre-Zhang Toushe image is calculated according to an optimization equation, and the pre-transmission image is updated.
Specifically, one of the two images is fixed, and the previous transmission image T is fixed in this embodiment 1 I.e. T 1 As a known quantity, the latter transmission image T is calculated according to an optimization equation 2 Is the optimal solution of (a), the optimization equation is:
E＝E d +λ s E s +λ t E t (1)
wherein T represents a transmission image, T 1 Representing a front transmission image, T 2 Representing rear Zhang Toushe image, E d For data fidelity terms, E s Is a space item, E t As a function of the time term,representing first order guidesThe number L represents the second derivative, F represents the two transmission images T 1 And T 2 Brightness relationship between (u) x (x,y),u y (x, y)) represents the deformation between images and E represents the total energy.
Will T 1 As a known quantity, combine the input image I 1 And I 2 By calculating the input image I 1 And I 2 Front transmission image T 1 Solving to minimize the total energy E in the optimization equation (i.e., calculating the first and second derivatives for each pixel point in the image) to obtain the optimal solution for the post Zhang Toushe image, and using the optimal solution as the new post Zhang Toushe image T 2 I.e. updated Zhang Toushe image T based on the optimal solution 2 。
Second, based on updated post Zhang Toushe image T 2 Fixed image T 2 Repeating the steps, and recalculating and solving to obtain a front transmission image T 1 Based on the best solution of Zhang Toushe image T after updating 1 。
In the iterative process, the optimization equation is used for solving and calculating, and the transmission images initialized by the two input images are updated, wherein the optimization equation not only considers the motion information between the two images, but also considers the space information in a single image, and meanwhile sparsely constrains the gradient of the reflecting layer between the two images, so that the space smoothness of the antireflection image is ensured, and the antireflection image with better effect is obtained.
Based on the above, step S4 is executed finally, that is, based on the two updated transmission images, the alignment and updating operations are executed continuously and circularly until the maximum number of the circulation iterations is reached, and the finally updated transmission image is output, that is, the final multiple anti-reflection images are output.
Specifically, in the current iteration process, two updated transmission images T are obtained through the calculation 1 And T 2 Judging whether the current iteration times beta is greater than or equal to the preset maximum iteration times beta max If yes, stopping iteration and outputting two transmission images T 1 And T 2 I.e. outputting the final anti-reflection image T 1 And T 2 The method comprises the steps of carrying out a first treatment on the surface of the Otherwise, if not, updating beta to beta+1, repeating the steps S2-S3, and performing cyclic iteration until the iteration number is greater than the preset maximum iteration number beta max 。
In the method for removing the light of the multiple images based on the space-time information, which is provided by the embodiment, not only the space information of each image is utilized, but also the motion information between the two images is utilized, and compared with a single image deblurring algorithm, more information in the multiple images is utilized, so that the light removing effect is better; by combining the para-LAP algorithm, even if brightness difference exists between two input images, the motion information between the images can be accurately estimated; the method provided by the embodiment does not depend on training data and high-performance computing equipment, does not need extra training time, and is economical and efficient.
Example two
The embodiment provides a multi-image antireflection system based on space-time information, which comprises:
the image acquisition and preprocessing module is used for acquiring an image sequence comprising a plurality of images, extracting front and rear images in the image sequence, taking the extracted two images as input images, and initializing a transmission image of the input images;
the image alignment module is used for calculating deformation between the two transmission images based on the transmission images initialized by the two input images by using a para-LAP algorithm to obtain motion information and brightness relation between the two transmission images, so as to align the two transmission images;
the image processing module is used for fixing the previous transmission image, calculating the optimal solution of the post Zhang Toushe image according to the optimization equation, and updating the post Zhang Toushe image; based on the updated post Zhang Toushe image, calculating a pre-Zhang Toushe image according to an optimization equation, and updating a pre-transmission image;
and the anti-reflection image output module is used for continuously and circularly executing the alignment and updating operation based on the two updated transmission images until the maximum number of the circular iterations is reached, and outputting the finally updated transmission images, namely outputting a plurality of final anti-reflection images.
Example III
The present embodiment provides an electronic device comprising a memory and a processor and computer instructions stored on the memory and running on the processor, which when executed by the processor, perform the steps in the spatio-temporal information based multi-image anti-reflection method described above.
Example IV
The present embodiment also provides a computer readable storage medium storing computer instructions that, when executed by a processor, perform the steps in the spatio-temporal information based multi-image anti-reflection method described above.
The steps involved in the second to fourth embodiments correspond to the first embodiment of the method, and the detailed description of the second embodiment refers to the relevant description of the first embodiment. The term "computer-readable storage medium" should be taken to include a single medium or multiple media including one or more sets of instructions; it should also be understood to include any medium capable of storing, encoding or carrying a set of instructions for execution by a processor and that cause the processor to perform any one of the methods of the present invention.
It will be appreciated by those skilled in the art that the modules or steps of the invention described above may be implemented by general-purpose computer means, alternatively they may be implemented by program code executable by computing means, whereby they may be stored in storage means for execution by computing means, or they may be made into individual integrated circuit modules separately, or a plurality of modules or steps in them may be made into a single integrated circuit module. The present invention is not limited to any specific combination of hardware and software.
While the present invention has been described in connection with the preferred embodiments, it should be understood that the present invention is not limited to the specific embodiments, but is set forth in the following claims.
Claims (10)
1. A method for removing light from a plurality of images based on spatio-temporal information, comprising:
acquiring an image sequence comprising a plurality of images, extracting a front image and a rear image in the image sequence, taking the extracted two images as input images, and initializing a transmission image of the input images;
calculating deformation between the two transmission images by using a para-LAP algorithm based on the transmission images initialized by the two input images to obtain motion information and brightness relation between the two transmission images, and further aligning the two transmission images;
fixing the front transmission image, calculating the optimal solution of the rear Zhang Toushe image according to an optimization equation, and updating the rear Zhang Toushe image; based on the updated post Zhang Toushe image, calculating an optimal solution of the pre-Zhang Toushe image according to an optimization equation, and updating the pre-transmission image;
based on the updated two transmission images, the alignment and updating operation is continuously and circularly executed until the maximum number of circulation iteration times is reached, and finally updated transmission images are output, namely, a plurality of final anti-reflection images are output.
2. The method for removing reflection from multiple images based on spatio-temporal information according to claim 1, wherein said initializing the transmission image of the input image means assigning the input image directly to the transmission image, i.e. input image I 1 As an initialized transmission image T 1 Input image I 2 As an initialized transmission image T 2 。
3. The spatio-temporal information based multi-image retroreflective method of claim 1, wherein said optimization equation is:
E＝E d +λ s E s +λ t E t
wherein T represents a transmission image, T 1 Representing a front transmission image, T 2 Representing the rear Zhang Toushe image, I 1 Representing a pre-tensioned input image, I 2 Representing post-tensioned input images, E d For data fidelity terms, E s Is a space item, E t As a function of the time term,represents the first derivative, L represents the second derivative, F represents the two transmission images T 1 And T 2 Brightness relationship between (u) x (x,y),u y (x, y)) represents the deformation between the front and rear transmission images, and E represents the total energy.
4. The method for spatio-temporal information based multi-image retroreflective sheeting of claim 3 wherein fixing the prior transmissive image, calculating an optimal solution for the post Zhang Toushe image according to an optimization equation, updating the post Zhang Toushe image, comprises:
the front transmission image T 1 As a known quantity, two front and rear images I of the input are combined 1 And I 2 Calculating the front and back images I of the input 1 And I 2 Front transmission image T 1 Solving for the first and second derivatives of (a) to minimize the total energy E in the optimization equation, solving for the optimal solution of the post Zhang Toushe image, and taking the optimal solution as the new post Zhang Toushe image T 2 。
5. The method for removing light from multiple images based on spatio-temporal information of claim 4, wherein calculating the first and second derivatives of the image means calculating the first and second derivatives of each pixel in the image.
6. A spatio-temporal information based multi-image retroreflective system comprising:
the image acquisition and preprocessing module is used for acquiring an image sequence comprising a plurality of images, extracting front and rear images in the image sequence, taking the extracted two images as input images, and initializing a transmission image of the input images;
the image alignment module is used for calculating deformation between the two transmission images based on the transmission images initialized by the two input images by using a para-LAP algorithm to obtain motion information and brightness relation between the two transmission images, so as to align the two transmission images;
the image processing module is used for fixing the previous transmission image, calculating the optimal solution of the post Zhang Toushe image according to the optimization equation, and updating the post Zhang Toushe image; based on the updated post Zhang Toushe image, calculating an optimal solution of the pre-Zhang Toushe image according to an optimization equation, and updating the pre-transmission image;
and the anti-reflection image output module is used for continuously and circularly executing the alignment and updating operation based on the two updated transmission images until the maximum number of the circular iterations is reached, and outputting the finally updated transmission images, namely outputting a plurality of final anti-reflection images.
7. The spatio-temporal information based multi-image retroreflective system of claim 6 wherein initializing a transmission image of an input image refers to assigning the input image directly to the transmission image, i.e., input image I 1 As an initialized transmission image T 1 Input image I 2 As an initialized transmission image T 2 。
8. The spatio-temporal information based multi-image retroreflective system of claim 6 wherein said optimization equation is:
E＝E d +λ s E s +λ t E t
wherein T represents a transmission image, T 1 Representing a front transmission image, T 2 Representing the rear Zhang Toushe image, I 1 Representing a pre-tensioned input image, I 2 Representing post-tensioned input images, E d For data fidelity terms, E s Is a space item, E t As a function of the time term,represents the first derivative, L represents the second derivative, F represents the two transmission images T 1 And T 2 Brightness relationship between (u) x (x,y),u y (x, y)) represents the deformation between the front and rear transmission images, and E represents the total energy.
9. An electronic device comprising a memory and a processor and computer instructions stored on the memory and running on the processor, which when executed by the processor, perform the steps of a spatio-temporal information based multi-image anti-reflection method of any of claims 1 to 5.
10. A computer readable storage medium storing computer instructions which, when executed by a processor, perform the steps of a spatio-temporal information based multi-image anti-reflection method of any of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311434745.4A CN117218040A (en) | 2023-10-31 | 2023-10-31 | Method and system for removing reflection of multiple images based on space-time information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311434745.4A CN117218040A (en) | 2023-10-31 | 2023-10-31 | Method and system for removing reflection of multiple images based on space-time information |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117218040A true CN117218040A (en) | 2023-12-12 |
Family
ID=89041167
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311434745.4A Pending CN117218040A (en) | 2023-10-31 | 2023-10-31 | Method and system for removing reflection of multiple images based on space-time information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117218040A (en) |
-
2023
- 2023-10-31 CN CN202311434745.4A patent/CN117218040A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108537746B (en) | Fuzzy variable image blind restoration method based on deep convolutional network | |
| Tai et al. | Richardson-lucy deblurring for scenes under a projective motion path | |
| CN112381897B (en) | Low-illumination image enhancement method based on self-coding network structure | |
| CN112435191B (en) | Low-illumination image enhancement method based on fusion of multiple neural network structures | |
| CN111915525A (en) | Low-illumination image enhancement method based on improved depth separable generation countermeasure network | |
| Vasluianu et al. | NTIRE 2023 image shadow removal challenge report | |
| CN111640060A (en) | Single image super-resolution reconstruction method based on deep learning and multi-scale residual dense module | |
| CN112102182A (en) | Single image reflection removing method based on deep learning | |
| US20220156891A1 (en) | Methods and systems for deblurring blurry images | |
| CN113570516A (en) | Image blind motion deblurring method based on CNN-Transformer hybrid self-encoder | |
| CN111461998A (en) | Environment reconstruction method and device | |
| CN110415193A (en) | The restored method of coal mine low-light (level) blurred picture | |
| Sun et al. | Deep maximum a posterior estimator for video denoising | |
| Mehta et al. | Gated multi-resolution transfer network for burst restoration and enhancement | |
| CN115810112A (en) | Image processing method, image processing device, storage medium and electronic equipment | |
| CN113628134A (en) | Image noise reduction method and device, electronic equipment and storage medium | |
| Sun et al. | MIPI 2023 Challenge on RGBW Remosaic: Methods and Results | |
| CN111861940A (en) | Image toning enhancement method based on condition continuous adjustment | |
| CN117218040A (en) | Method and system for removing reflection of multiple images based on space-time information | |
| Zhang et al. | Single Image Reflection Removal Based on Dark Channel Sparsity Prior | |
| Yang et al. | MCFD: A hardware-efficient noniterative multicue fusion demosaicing algorithm | |
| Li et al. | Rendering nighttime image via cascaded color and brightness compensation | |
| CN115311149A (en) | Image denoising method, model, computer-readable storage medium and terminal device | |
| Barua et al. | Arthdr-net: Perceptually realistic and accurate hdr content creation | |
| Zhang et al. | Attention driven self-similarity capture for motion deblurring |
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 |