CN114745531A - Image white balance correction method and related device - Google Patents
Image white balance correction method and related device Download PDFInfo
- Publication number
- CN114745531A CN114745531A CN202210336143.4A CN202210336143A CN114745531A CN 114745531 A CN114745531 A CN 114745531A CN 202210336143 A CN202210336143 A CN 202210336143A CN 114745531 A CN114745531 A CN 114745531A
- Authority
- CN
- China
- Prior art keywords
- image
- white balance
- images
- balance correction
- light sources
- 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
- 238000012937 correction Methods 0.000 title claims abstract description 133
- 238000000034 method Methods 0.000 title claims abstract description 69
- 238000013135 deep learning Methods 0.000 claims abstract description 51
- 238000009877 rendering Methods 0.000 claims abstract description 47
- 238000012549 training Methods 0.000 claims description 33
- 238000013136 deep learning model Methods 0.000 claims description 31
- 238000007499 fusion processing Methods 0.000 claims description 18
- 238000004590 computer program Methods 0.000 claims description 11
- 238000012545 processing Methods 0.000 abstract description 25
- 230000008569 process Effects 0.000 abstract description 15
- 238000013527 convolutional neural network Methods 0.000 description 13
- 230000006870 function Effects 0.000 description 10
- 238000010586 diagram Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 4
- 238000000605 extraction Methods 0.000 description 3
- 230000007935 neutral effect Effects 0.000 description 2
- 230000005457 Black-body radiation Effects 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000002156 mixing Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- WFKWXMTUELFFGS-UHFFFAOYSA-N tungsten Chemical compound [W] WFKWXMTUELFFGS-UHFFFAOYSA-N 0.000 description 1
- 229910052721 tungsten Inorganic materials 0.000 description 1
- 239000010937 tungsten Substances 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N9/00—Details of colour television systems
- H04N9/64—Circuits for processing colour signals
- H04N9/73—Colour balance circuits, e.g. white balance circuits or colour temperature control
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/80—Camera processing pipelines; Components thereof
- H04N23/84—Camera processing pipelines; Components thereof for processing colour signals
- H04N23/88—Camera processing pipelines; Components thereof for processing colour signals for colour balance, e.g. white-balance circuits or colour temperature control
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N9/00—Details of colour television systems
- H04N9/64—Circuits for processing colour signals
- H04N9/646—Circuits for processing colour signals for image enhancement, e.g. vertical detail restoration, cross-colour elimination, contour correction, chrominance trapping filters
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Processing Of Color Television Signals (AREA)
- Color Television Image Signal Generators (AREA)
Abstract
The embodiment of the application discloses an image white balance correction method and a related device. The image white balance correction method comprises the following steps: acquiring an image to be calibrated; rendering the image to be calibrated based on N kinds of preset ambient light sources to obtain N first images; the first images correspond to preset ambient light sources one to one, and N is an integer greater than 1; and inputting the N first images into a deep learning white balance correction model to obtain a white balance correction image corresponding to the image to be calibrated. By the mode, the mixed color temperature scene does not need to be judged in advance, so that the flow of white balance processing on the image is simplified; and the estimation error of the ambient light source is not introduced in the process of carrying out white balance processing on the mixed color temperature image, so that the accuracy of the white balance processing is improved.
Description
Technical Field
The present application relates to the field of information processing, and in particular, to a method and a related apparatus for correcting white balance of an image.
Background
In order to solve the problem of imaging color cast and reduce the influence of the color of an ambient light source, the color of an object is accurately reflected as much as possible when a camera images, and a white balance correction algorithm is provided. The traditional white balance correction algorithm can achieve better effect on an image (hereinafter referred to as a monochromatic temperature image) under a scene with a single color temperature light source. However, in practical situations, two or more color temperature light sources (hereinafter referred to as mixed color temperature scenes) are usually present, and in such mixed color temperature scenes, it is difficult for the conventional white balance correction algorithm to achieve a good effect.
In order to solve the problem that the traditional white balance correction algorithm has poor effect in a mixed color temperature scene, generally, after the image to be processed is judged to be the image in the mixed color temperature scene, a plurality of ambient light sources in the mixed color temperature scene are respectively estimated, and each color temperature scene in the mixed color temperature scene is respectively subjected to white balance processing according to the estimation result.
However, with such a white balance processing method, it is possible to introduce an error in estimation of the ambient light source, thereby reducing the accuracy of white balance processing on an image in a mixed color temperature scene (hereinafter, referred to as a mixed color temperature image). Therefore, how to improve the accuracy of image white balance processing in a mixed color temperature scene is an urgent problem to be solved.
Disclosure of Invention
The embodiment of the application provides an image white balance correction method and a related device, through the method, in the process of carrying out white balance processing on a mixed color temperature image, estimation errors of an ambient light source cannot be introduced, deep semantic feature learning and white balance processing are carried out through a deep learning model, and the accuracy of the white balance processing on the mixed color temperature image can be improved.
In a first aspect, an embodiment of the present application provides an image white balance correction method, including:
acquiring an image to be calibrated; rendering the image to be calibrated based on N kinds of preset environment light sources to obtain N pieces of first images; the first images correspond to preset ambient light sources one to one, and N is an integer greater than 1; and inputting the N first images into the deep learning white balance correction model to obtain a white balance correction image corresponding to the image to be calibrated.
Therefore, the mixed color temperature image is corrected by the image white balance correction method, so that the introduction of estimation errors of the mixed color temperature environment light source can be avoided, deep semantic feature extraction and white balance correction are carried out on the basis of the deep learning model, and the white balance processing accuracy of the mixed color temperature image can be remarkably improved.
In one possible implementation, the image to be calibrated includes a single color temperature image or a mixed color temperature image. By implementing the possible implementation mode, the judgment of the mixed color temperature scene is not required to be carried out in advance before the white balance processing is carried out on the image to be calibrated, so that the flow of the white balance processing on the image is simplified.
In a possible implementation manner, inputting N first images into a deep learning white balance correction model to obtain N first weight images; the first weight images correspond to the first images one to one; and carrying out fusion processing on the N first images based on the N first weight images to obtain a white balance correction image corresponding to the image to be calibrated.
In one possible implementation mode, a white balance label image corresponding to each single-color temperature image in a plurality of single-color temperature images is obtained; rendering each white balance label image based on M kinds of preset environment light sources to obtain a mixed color temperature image corresponding to each white balance label image; wherein M is an integer greater than 1 and less than or equal to N; and training a preset deep learning model based on each white balance label image and the corresponding mixed color temperature image to obtain a deep learning white balance correction model.
In one possible implementation, a plurality of monochromatic temperature images are acquired; wherein each monochromatic temperature image is accompanied by a color chip; determining a light source estimation value corresponding to each monochromatic temperature image based on a color chart in each monochromatic temperature image; and carrying out white balance correction on each monochromatic temperature image based on the light source estimation value of each monochromatic temperature image to obtain a white balance label image corresponding to each monochromatic temperature image.
In one possible implementation manner, each white balance label image is rendered based on M kinds of preset ambient light sources to obtain M second images corresponding to each white balance label image; the second images correspond to preset ambient light sources one to one; acquiring M second weight images corresponding to each white balance label image; and performing fusion processing on the M second images based on the M second weight images to obtain a mixed color temperature image corresponding to each white balance label image.
In one possible implementation manner, rendering is performed on the mixed color temperature image corresponding to each white balance label image based on N kinds of preset ambient light sources to obtain N third images; the third image corresponds to a preset environment light source one by one; and training a preset deep learning model based on the N third images and the corresponding white balance label images to obtain a deep learning white balance correction model.
In a second aspect, an embodiment of the present application provides an image white balance correction apparatus including:
the image acquisition unit is used for acquiring an image to be calibrated;
the image rendering unit is used for rendering the image to be calibrated based on the N types of preset environment light sources to obtain N first images; the first images correspond to preset ambient light sources one to one, and N is an integer greater than 1;
and the white balance correction unit is used for inputting the N first images into the deep learning white balance correction model to obtain a white balance correction image corresponding to the image to be calibrated.
In one possible implementation, the image to be calibrated includes a single color temperature image or a mixed color temperature image.
In one possible implementation, the white balance correction unit is further configured to: inputting N first images into a deep learning white balance correction model to obtain N first weight images; the first weight images correspond to the first images one to one; and carrying out fusion processing on the N first images based on the N first weight images to obtain a white balance correction image corresponding to the image to be calibrated.
In one possible implementation manner, inputting the N first images into the deep learning white balance correction model to obtain a white balance correction image corresponding to the image to be calibrated, and the image obtaining unit is further configured to obtain a white balance label image corresponding to each of the plurality of monochromatic temperature images; the image rendering unit is further used for rendering each white balance label image based on M kinds of preset ambient light sources to obtain a mixed color temperature image corresponding to each white balance label image; wherein M is an integer greater than 1 and less than or equal to N;
in a possible implementation manner, the device further includes a training unit, and the training unit is configured to train a preset deep learning model based on each white balance label image and the corresponding mixed color temperature image, so as to obtain a deep learning white balance correction model.
In a possible implementation manner, the image obtaining unit is further configured to: acquiring a plurality of monochromatic temperature images; wherein each monochromatic temperature image is attached with a color chip; determining a light source estimation value of each monochromatic temperature image based on a color chart in each monochromatic temperature image; and carrying out white balance correction on each monochromatic temperature image based on the light source estimation value of each monochromatic temperature image to obtain a white balance label image corresponding to each monochromatic temperature image.
In a possible implementation manner, the image rendering unit is further configured to: rendering each white balance label image based on M kinds of preset environment light sources to obtain M second images corresponding to each white balance label image; the second images correspond to preset ambient light sources one to one; acquiring M second weight images corresponding to each white balance label image; and performing fusion processing on the M second images based on the M second weight images to obtain a mixed color temperature image corresponding to each white balance label image.
In a possible implementation manner, the image rendering unit is further configured to render the mixed color temperature image corresponding to each white balance label image based on N preset ambient light sources to obtain N third images; the third image corresponds to a preset environment light source one by one; and the training unit is also used for training a preset deep learning model based on the N third images and the corresponding white balance label images to obtain a deep learning white balance correction model.
In a third aspect, the present application provides a chip, where the chip is configured to obtain an image to be calibrated, where the image to be calibrated is a monochromatic temperature image or a mixed color temperature image; rendering the image to be calibrated based on N kinds of preset environment light sources to obtain N pieces of first images; the first images correspond to preset ambient light sources one by one, and N is an integer greater than 1; and inputting the N first images into the deep learning white balance correction model to obtain a white balance correction image corresponding to the image to be calibrated.
In a fourth aspect, the present application provides a chip module, where the chip die holder includes the chip in the third aspect.
In a fifth aspect, the present application provides a computer device comprising:
a memory for storing a computer program;
a processor invoking a computer program for performing the following operations: acquiring an image to be calibrated, wherein the image to be calibrated is a single-color temperature image or a mixed color temperature image; rendering the image to be calibrated based on N kinds of preset environment light sources to obtain N pieces of first images; the first images correspond to preset ambient light sources one to one, and N is an integer greater than 1; and inputting the N first images into the deep learning white balance correction model to obtain a white balance correction image corresponding to the image to be calibrated.
In a sixth aspect, an embodiment of the present application provides a computer-readable storage medium for storing computer software instructions for the computer device, which includes a program for executing the method according to any one of the first aspect.
Drawings
Fig. 1 is a schematic flowchart of an image white balance correction method according to an embodiment of the present disclosure;
fig. 2 is a schematic flowchart of another image white balance correction method according to an embodiment of the present disclosure;
FIG. 3a is a schematic diagram of a model training process according to an embodiment of the present application;
FIG. 3b is a schematic diagram of another model training process provided in the embodiments of the present application;
fig. 4 is a schematic structural diagram of an image white balance correction apparatus according to an embodiment of the present disclosure;
fig. 5 is a schematic structural diagram of a computer device according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application clearer, the present application will be described in further detail with reference to the accompanying drawings.
The terms "first" and "second", etc. in the description, claims and drawings of the present application are used for distinguishing between different objects and not for describing a particular order. Furthermore, the terms "include" and "have," as well as any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, article, or apparatus that comprises a list of operations or elements is not limited to those listed but may alternatively include other operations or elements not listed or inherent to such process, method, article, or apparatus.
Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. It is explicitly and implicitly understood by one skilled in the art that the embodiments described herein can be combined with other embodiments.
In the present application, "at least one" means one or more, "a plurality" means two or more, "at least two" means two or three and three or more, "and/or" for describing the correspondence of the corresponding objects, indicating that three relationships may exist, for example, "a and/or B" may indicate: only A, only B and both A and B are present, wherein A and B may be singular or plural. The character "/" generally indicates that the preceding and following corresponding pair is in an "or" relationship. "at least one of the following" or similar expressions refer to any combination of these items, including any combination of single item(s) or plural items. For example, at least one (one) of a, b, or c, may represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", wherein a, b, c may be single or plural.
The image white balance correction method in the present application can be applied to a computer device, and it should be understood that the computer device mentioned in the present application includes a terminal device or a server. To facilitate an understanding of the embodiments disclosed herein, some concepts related to the embodiments of the present application will be first explained. The description of these concepts includes, but is not limited to, the following.
1. Color temperature
The color of a light source is often expressed in terms of color temperature. The color temperature of a light source is defined as the kelvin temperature of a black body radiator that emits light similar to that of the light source. Color temperature is of great significance in the fields of photography, video recording, publishing and the like. The color temperature of the light source is different, and the feeling brought by the color temperature is different.
The color of light emitted by the light source is the same as the color of light radiated by the black body at a certain temperature, and the temperature of the black body is the color temperature of the light source. In black body radiation, the color of light is different with different temperatures, and the black body presents a gradual change process of red, orange red, yellow-white, white and blue-white.
2. White balance correction
The white balance correction process can be understood as: under any light source, the image processing process of color restoration is carried out on the image of the white object which is originally made of the material. After the white balance of the image is corrected, the influence of the color temperature of the external light source on the image can be removed, so that an object which is originally made of white can be displayed as white on a photo.
3. Color card
A color card, also called color reference card, is a flat physical object with many different color samples. The color chart may be provided in the form of a single-page image table, or may be provided in the form of a color plate. It should be understood that the color cards referred to in this application include, but are not limited to, 24 color cards, gray cards, and the like.
In order to better understand the scheme provided by the present application, the following will explain the embodiments of the present application by referring to the attached images in the embodiments of the present application.
Referring to fig. 1, fig. 1 is a schematic flow chart of an image white balance correction method according to an embodiment of the present disclosure. The execution subject of the method shown in fig. 1 may be a computer device, or a chip in the computer device. Fig. 1 illustrates an example of an execution subject of the method by using a computer device. As shown in fig. 1, the image white balance correction method includes steps S101 to S103.
And S101, acquiring an image to be calibrated.
The computer equipment for white balance processing can acquire an image to be calibrated through an image acquisition module of the computer equipment. Alternatively, the computer device may also receive images to be calibrated from other devices, such as an image acquisition device or other devices having white balance processing requirements. The image to be calibrated includes one or more of a single color temperature image or a mixed color temperature image, and the number of the image to be calibrated may be one or more, which is not specifically limited in this application. It is to be understood that the mixed color temperature image may be understood as an image under various color temperatures.
S102, rendering the image to be calibrated based on the N preset environment light sources to obtain N first images. The first images correspond to preset ambient light sources one to one, and N is an integer greater than 1.
In other words, a plurality of types of ambient light sources are preset, each type of preset ambient light source is used to render the image to be calibrated, and an image rendered by each type of preset ambient light source (i.e., the first image) is obtained. Specifically, for each preset ambient light source, a color gain value (Rgain, Ggain, Bgain) corresponding to each preset ambient light source may be calculated, and the color gain value is applied to three RGB channels of the image to be calibrated, so as to generate the image to be calibrated (i.e., the first image) rendered by each preset ambient light source.
The method for presetting the plurality of ambient light sources includes, but is not limited to, the following methods:
the first method is as follows: selecting a plurality of color temperature light sources according to a preset color temperature range, for example, the preset color temperature range is 1000 k-15000 k, 1000k is used as a first preset color temperature light source, and one color temperature light source is selected every 1000k, so as to obtain 15 kinds of preset color temperature light sources: 1000k, 2000k, 3000k, … …, 14000k, 15000 k.
The second method comprises the following steps: the color temperature light sources are set according to common laboratory light source types, such as color temperature light sources set as D75, D65, D50, TL84, CWF, TL83, A and the like.
The third method comprises the following steps: collecting a plurality of monochromatic temperature images with color cards, and acquiring a light source estimation value calibrated by each monochromatic temperature image; clustering the light source estimation values of the multiple single color temperature images by using a clustering algorithm (such as a kmeans clustering algorithm) to obtain a clustering result of the multiple light source estimation values; and selecting the light source estimation value corresponding to each clustering center as a preset environment light source.
S103, inputting the N first images into a deep learning white balance correction model to obtain a white balance correction image corresponding to the image to be calibrated.
The deep learning white balance correction model includes, but is not limited to, a Convolutional Neural Network (CNN), a Recurrent Neural Network (RNN), a Transformer, and the like.
In one embodiment, the N first images are input into a deep learning white balance correction model to obtain N first weight images corresponding to a preset ambient light source, wherein the first weight images correspond to the first images one to one; and performing fusion processing on the N first images based on the N first weight images to obtain a white balance correction image corresponding to the image to be calibrated.
For example, after rendering the image to be calibrated based on 4 preset ambient light sources, 4 first images are obtained: p1, P2, P3 and P4. The 4 first images are input into a depth learning white balance correction model to obtain first weight images corresponding to each first image, wherein the first weight image corresponding to P1 is W1, the first weight image corresponding to P2 is W2, the first weight image corresponding to P3 is W3, and the first weight image corresponding to P4 is W4. The 4 first images are fused based on the 4 first weighted images, and the white balance correction image corresponding to the image to be calibrated is obtained as (P1 × W1+ P2 × W2+ P3 × W3+ P4 × W4)/(W1+ W2+ W3+ W4).
In summary, compared with the method of separately performing white balance processing on each color temperature according to the estimation result after estimating the ambient light source of the mixed color temperature scene, the white balance calibration method provided in fig. 1 does not need to perform judgment on the mixed color temperature scene in advance, and is suitable for all types of images to be processed, so that the white balance processing flow of the images is simplified; and the estimation error of the ambient light source is not introduced in the process of carrying out white balance processing on the mixed color temperature image, and deep semantic feature extraction and white balance correction are carried out on the basis of the deep learning white balance correction model, so that the accuracy of white balance processing is improved.
After the white balance calibration method is introduced, how to obtain the deep learning white balance correction model is described in detail below. Referring to fig. 2, fig. 2 is a schematic flow chart of another image white balance correction method according to an embodiment of the present application. The execution subject of the method shown in fig. 2 may be a computer device, or a chip in the computer device. Fig. 2 illustrates an example of an execution subject of the method by using a computer device. As shown in fig. 2, the image white balance correction method includes steps S201 to S206.
S201, acquiring a white balance label image corresponding to each single-color temperature image in the multiple single-color temperature images.
The white balance label image is used for carrying out model training on a preset deep learning model to obtain a deep learning white balance correction model.
In one possible embodiment, the manner of acquiring the white balance label image is: acquiring a plurality of monochromatic temperature images; wherein each monochromatic temperature image is accompanied by a color chip. And further, performing white balance correction processing on each monochromatic temperature image based on the light source estimation value of each monochromatic temperature image to obtain a white balance label image (namely an image after white balance correction corresponding to each monochromatic temperature image) corresponding to each monochromatic temperature image.
In a specific example, the manner of acquiring the white balance tag image includes: a plurality of monochromatic temperature images are collected in different background environments, different color temperatures and different brightness environments, so that the diversity of sample data is increased. The collecting light source includes, but is not limited to, laboratory artificial light sources such as a, H, TL83, TL84, D50, D65 and the like, fluorescent lamps, tungsten lamps, and outdoor daylight in the morning, in the afternoon and in the evening. Each monochromatic temperature image is acquired with a color card in the image, and the color card is used for calculating the true light source value of the monochromatic temperature image through an achromatic (neutral) area. After a plurality of monochromatic temperature images are collected, the positions of the color cards and the neutral color blocks are marked by using a color card detection algorithm or a manual marking method and the like. For example, taking a 24-color card as an example, color patches 19-24 of the color card are used as achromatic color patches, the brightest color patch without overexposure is used as a true value calibration area, and the median of the color patch is used as a true value of the light source. Further, based on a light source true value obtained by a color card eliminating block area in the collected image, a global color gain value corresponding to each single-color temperature image is calculated, and based on the global color gain value of each single-color temperature image, white balance correction processing is performed on each single-color temperature image, so that a corresponding white balance label image after each single-color temperature image is subjected to accurate white balance correction is obtained. It should be understood that the white balance label image mentioned in the present application may also be referred to as a correct label (GT) image.
In one possible embodiment, in order to accelerate the subsequent processing speed and the training speed of the deep learning white balance correction model, the size of each white balance label image may be reduced to a uniform size.
S202, rendering each white balance label image based on M kinds of preset ambient light sources to obtain a mixed color temperature image corresponding to each white balance label image. Wherein M is an integer greater than 1 and less than or equal to N.
It can be understood that M kinds of preset ambient light sources are determined from the N kinds of preset ambient light sources shown in the foregoing S102, and each white balance label image is rendered based on the M kinds of preset ambient light sources, so as to obtain a mixed color temperature image corresponding to the white balance label image. For any white balance label image, the M kinds of preset ambient light sources corresponding to the white balance label image may be randomly selected from the N kinds of preset ambient light sources (may be randomly generated according to a random algorithm, and may be understood as randomly determining a plurality of kinds of preset ambient light sources from the N kinds of preset ambient light sources. It should be understood that the number of the preset ambient light sources corresponding to each white balance label image may be different (i.e., M may be different), and the type of the preset ambient light sources may also be different.
For example, 3 GT images are obtained in S201: GT image 1, GT image 2 and GT image 3, there are 15 preset ambient light sources in S102. Further, 3 preset ambient light sources may be randomly determined from the 15 preset ambient light sources, the GT image 1 is rendered based on the 3 preset ambient light sources, 5 preset ambient light sources are randomly determined from the 15 preset ambient light sources, the GT image 2 is rendered based on the 5 preset ambient light sources, 10 preset ambient light sources are randomly determined from the 15 preset ambient light sources, and the GT image 3 is rendered based on the 10 preset ambient light sources.
In a possible implementation manner, each white balance label image is rendered based on M preset ambient light sources to obtain a mixed color temperature image corresponding to each white balance label image, and the specific implementation manner may be: rendering each white balance label image based on M kinds of preset environment light sources to obtain M second images corresponding to each white balance label image; the second images correspond to preset ambient light sources one by one; acquiring M second weight images corresponding to each white balance label image; and performing fusion processing on the M second images based on the M second weighted images to obtain a mixed color temperature image corresponding to each white balance label image.
For each white balance label image, the operation step of obtaining the corresponding mixed color temperature image comprises the following steps: rendering the white balance label image by using each preset environment light source in the M preset environment light sources respectively, and acquiring a corresponding rendering image (namely a second image) under each preset environment light source. In particular, it may be possible to target each kindAnd presetting the ambient light sources, calculating corresponding color gain values (Rgain, Ggain, Bgain) of the ambient light sources, and applying the color gain values to the RGB three channels of the white balance label image to generate a corresponding rendering image slice (namely a second image) under each preset ambient light source. Further, for each preset ambient light source, a second weight image is randomly generated in accordance with the size of each white balance label image. The manner of randomly generating the second weighted image includes, but is not limited to: the auxiliary generation mode including semantic information is based on certain distribution such as polynomial distribution and exponential distribution or mixed distribution of several kinds, and is based on heatmap and depthmap. In one possible implementation, after the M second weight images are generated, each second weight image needs to be normalized pixel by pixel. Further, performing fusion processing on the M second images and the M second weights of each white balance label image to generate a mixed color temperature image corresponding to the white balance label image. Exemplarily, for the k-th white balance label image, its corresponding mixed color temperature image DkAs shown in equation (1).
Wherein D iskC is a c-th preset ambient light source among the M preset ambient light sources,
second image rendered for the c-th preset ambient light source
The corresponding second weight image.
S203, training a preset deep learning model based on each white balance label image and the corresponding mixed color temperature image to obtain a deep learning white balance correction model. The preset deep learning model includes, but is not limited to, CNN, RNN, transform, and the like.
The color temperature image and the white balance label image are mixed and input into the preset deep learning model, so that the preset deep learning model can continuously adjust parameters according to the input image and the white balance label image corresponding to the input image to obtain the deep learning white balance correction model. The predetermined deep learning model includes, but is not limited to, CNN, RNN, transform, etc.
In one possible embodiment, the specific operation of training the preset deep learning model to obtain the deep learning white balance correction model is as follows: rendering the mixed color temperature image corresponding to each white balance label image based on N kinds of preset environment light sources to obtain N third images; and the third image corresponds to the preset ambient light sources one to one. And further, training a preset deep learning model based on the N third images and the corresponding white balance label images as the input of the preset deep learning model to obtain the deep learning white balance correction model. In the following, a preset deep learning model is taken as an example of the CNN network, and a training process is briefly illustrated. In one example, as shown in fig. 3a, the plurality of third images are input into a CNN network model, and after being processed by the CNN network model, a white balance correction image corresponding to the mixed color temperature image is obtained. In another example, as shown in fig. 3b, the preset deep learning model includes a generator network and a discriminator network, and the plurality of third images are input into the preset deep learning model. It is to be understood that, in the case where the preset deep learning model is as shown in fig. 3b, the purpose of the discriminator network is to distinguish the output result of the generation network (i.e., the white balance correction image) from the white balance label image; the purpose of the generator network is to generate a white balance corrected image from the mixed color temperature image and to make it impossible for the discriminator to distinguish between the white balance corrected image and the white balance label image, thereby causing the generator network to generate a more effective white balance corrected image.
The CNN network shown in fig. 3a is specifically described below with a preset deep learning model, where the CNN network includes a series of convolutional layers and activation function layers. And for the mixed color temperature image of each white balance label image, rendering the mixed color temperature image based on the N preset environment light sources to obtain N third images, wherein the third images correspond to the preset environment light sources one by one. Further, the N images are input into the CNN network. And performing feature extraction through a series of convolutional layer operations in the CNN network, and inputting the extracted features into a softmax layer across channel dimensions to obtain a third weight image with the same size as each third image. Furthermore, each third image and the corresponding third weight image are subjected to fusion processing, and a white balance correction image corresponding to the mixed color temperature image is output. Further, calculating a fitting Loss (Loss) function based on the white balance correction image and the white balance label image, and performing reverse training and updating on model parameters of the CNN network based on the calculated Loss until the CNN network converges. The Loss function is any Loss function that can measure the image reconstruction error, for example, the reconstructed Loss function can be shown in formula (2).
Loss＝‖GT-∑iMi⊙Ii‖norm (2)
Wherein |)normIs a norm, such as L1 norm, Frobenius norm, etc., IiIs N third images, MiN corresponding third weight images, which is a certain blending operation, such as a dot product operation. It should be noted that, in the training process of the model, besides the above reconstructed Loss function, a regularized Loss function may be introduced to prevent network overfitting, where the regularized Loss function includes regularization of model parameters, regularization of fusion weight images, and the like.
And S204, acquiring an image to be calibrated.
S205, rendering the image to be calibrated based on the N preset environment light sources to obtain N first images. The first images correspond to preset ambient light sources one to one, and N is an integer greater than 1.
S206, inputting the N first images into a deep learning white balance correction model to obtain a white balance correction image corresponding to the image to be calibrated.
The specific implementation manners of S204-S206 can refer to the foregoing descriptions of the specific implementation manners of S101-S103, and the repeated descriptions are omitted here.
Therefore, through the model training mode, the computer equipment can generate the deep learning white balance correction model according to the white balance label image and the mixed color temperature image, so that the accuracy of the white balance correction model can be improved, and the effect of white balance processing on the image is improved.
Referring to fig. 4, fig. 4 is a schematic structural diagram of an image white balance correction apparatus according to an embodiment of the present invention, the image white balance correction apparatus is configured in a computer device, and the image white balance correction apparatus includes:
an image acquisition unit 401, configured to acquire an image to be calibrated;
an image rendering unit 402, configured to render the image to be calibrated based on N preset ambient light sources, so as to obtain N first images; the first images correspond to preset ambient light sources one to one, and N is an integer greater than 1;
the white balance correction unit 403 is configured to input the N first images into the deep learning white balance correction model, and obtain a white balance correction image corresponding to the image to be calibrated.
In one possible implementation, the image to be calibrated includes a single color temperature image or a mixed color temperature image.
In one possible implementation manner, the white balance correction unit 403 is further configured to: inputting N first images into a deep learning white balance correction model to obtain N first weight images; the first weight images correspond to the first images one to one; and carrying out fusion processing on the N first images based on the N first weight images to obtain a white balance correction image corresponding to the image to be calibrated.
In a possible implementation manner, before inputting N first images into the deep learning white balance correction model to obtain a white balance correction image corresponding to an image to be calibrated, the image obtaining unit 401 is further configured to obtain a white balance label image corresponding to each of the plurality of monochromatic temperature images; the image rendering unit 402 is further configured to render each white balance label image based on M types of preset ambient light sources to obtain a mixed color temperature image corresponding to each white balance label image; wherein M is an integer greater than 1 and less than or equal to N;
in a possible implementation manner, the device further includes a training unit, and the training unit is configured to train a preset deep learning model based on each white balance label image and the corresponding mixed color temperature image, so as to obtain a deep learning white balance correction model.
In a possible implementation manner, the image obtaining unit 401 is further configured to: acquiring a plurality of monochromatic temperature images; wherein each monochromatic temperature image is accompanied by a color chip; determining a light source estimation value of each monochromatic temperature image based on a color chart in each monochromatic temperature image; and carrying out white balance correction processing on each monochromatic temperature image based on the light source estimation value of each monochromatic temperature image to obtain a white balance label image corresponding to each monochromatic temperature image.
In a possible implementation manner, the image rendering unit 402 is further configured to: rendering each white balance label image based on M kinds of preset environment light sources to obtain M second images corresponding to each white balance label image; the second images correspond to preset ambient light sources one to one; acquiring M second weight images corresponding to each white balance label image; and performing fusion processing on the M second images based on the M second weight images to obtain a mixed color temperature image corresponding to each white balance label image.
In a possible implementation manner, the image rendering unit 402 is further configured to render the mixed color temperature image corresponding to each white balance tag image based on N preset ambient light sources, so as to obtain N third images; the third image corresponds to a preset environment light source one by one; and the training unit is also used for training a preset deep learning model based on the N third images and the corresponding white balance label images to obtain a deep learning white balance correction model.
It should be noted that the functions of each unit module of the image white balance correction apparatus described in the embodiment of the present invention may be specifically implemented according to the method in the method embodiment of fig. 1 or fig. 2, and the specific implementation process may refer to the description related to the method embodiment of fig. 1 or fig. 2, which is not described herein again.
The embodiment of the present application further provides a chip, where the chip may perform relevant steps of the computer device in the foregoing method embodiments. The chip is used for: acquiring an image to be calibrated, wherein the image to be calibrated is a single-color temperature image or a mixed color temperature image; rendering the image to be calibrated based on N kinds of preset environment light sources to obtain N pieces of first images; the first images correspond to preset ambient light sources one by one, and N is an integer greater than 1; and inputting the N first images into the deep learning white balance correction model to obtain a white balance correction image corresponding to the image to be calibrated.
In one possible implementation, the chip is configured to: inputting N first images into a deep learning white balance correction model to obtain N first weight images; the first weight images correspond to the first images one to one; and carrying out fusion processing on the N first images based on the N first weight images to obtain a white balance correction image corresponding to the image to be calibrated.
In a possible implementation manner, the image to be calibrated includes a single color temperature image and a mixed color temperature image.
In one possible implementation, the chip is configured to: acquiring a white balance label image corresponding to each monochromatic temperature image in a plurality of monochromatic temperature images; rendering each white balance label image based on M kinds of preset environment light sources to obtain a mixed color temperature image corresponding to each white balance label image; wherein M is an integer greater than 1 and less than or equal to N; and training a preset deep learning model based on each white balance label image and the corresponding mixed color temperature image to obtain a deep learning white balance correction model.
In one possible implementation, the chip is configured to: acquiring a plurality of monochromatic temperature images; wherein each monochromatic temperature image is attached with a color chip; determining a light source estimation value of each monochromatic temperature image based on a color chart in each monochromatic temperature image; and carrying out white balance correction on each monochromatic temperature image based on the light source estimation value of each monochromatic temperature image to obtain a white balance label image corresponding to each monochromatic temperature image.
In one possible implementation, the chip is configured to: rendering each white balance label image based on M kinds of preset environment light sources to obtain M second images corresponding to each white balance label image; the second images correspond to preset ambient light sources one to one; acquiring M second weight images corresponding to each white balance label image; and performing fusion processing on the M second images based on the M second weight images to obtain a mixed color temperature image corresponding to each white balance label image.
In one possible implementation, the chip is configured to: rendering a mixed color temperature image corresponding to each white balance label image based on N kinds of preset ambient light sources to obtain N third images; the third image corresponds to a preset environment light source one by one; and training a preset deep learning model based on the N third images and the corresponding white balance label images to obtain a deep learning white balance correction model.
The embodiment of the present application further provides a chip module, which can be applied to a computer device, and the chip module includes the above chip that can be applied to the computer device.
Referring to fig. 5, fig. 5 is a schematic structural diagram of a computer device according to an embodiment of the present disclosure. The computer device 50 described in the embodiment of the present application includes: the processor 501, the memory 502, the processor 501 and the memory 502 are connected by one or more communication buses.
The Processor 501 may be a Central Processing Unit (CPU), and may also be other general purpose processors, Digital Signal Processors (DSP), Application Specific Integrated Circuits (ASIC), Field-Programmable Gate arrays (FPGA) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components, and so on. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The processor 501 is configured to support the user equipment to perform the corresponding functions of the computer device in the method of fig. 1 or fig. 2.
The memory 502 may include read-only memory and random access memory, and provides computer programs and data to the processor 501. A portion of the memory 502 may also include non-volatile random access memory. The processor 501, when calling the computer program, is configured to perform: acquiring an image to be calibrated; rendering the image to be calibrated based on N kinds of preset environment light sources to obtain N pieces of first images; the first images correspond to preset ambient light sources one to one, and N is an integer greater than 1; and inputting the N first images into the deep learning white balance correction model to obtain a white balance correction image corresponding to the image to be calibrated.
In one possible implementation, the image to be calibrated includes a single color temperature image or a mixed color temperature image.
In a possible implementation manner, inputting N first images into a deep learning white balance correction model to obtain N first weight images; the first weight images correspond to the first images one to one; and carrying out fusion processing on the N first images based on the N first weight images to obtain a white balance correction image corresponding to the image to be calibrated.
In one possible implementation mode, a white balance label image corresponding to each single-color temperature image in a plurality of single-color temperature images is obtained; rendering each white balance label image based on M kinds of preset environment light sources to obtain a mixed color temperature image corresponding to each white balance label image; wherein M is an integer greater than 1 and less than or equal to N; and training a preset deep learning model based on each white balance label image and the corresponding mixed color temperature image to obtain a deep learning white balance correction model.
In one possible implementation, a plurality of monochromatic temperature images are acquired; wherein each monochromatic temperature image is accompanied by a color chip; determining a light source estimation value of each monochromatic temperature image based on a color chart in each monochromatic temperature image; and carrying out white balance correction on each monochromatic temperature image based on the light source estimation value of each monochromatic temperature image to obtain a white balance label image corresponding to each monochromatic temperature image.
In one possible implementation manner, each white balance label image is rendered based on M kinds of preset ambient light sources to obtain M second images corresponding to each white balance label image; the second images correspond to preset ambient light sources one by one; acquiring M second weight images corresponding to each white balance label image; and performing fusion processing on the M second images based on the M second weight images to obtain a mixed color temperature image corresponding to each white balance label image.
In a possible implementation manner, rendering is performed on the mixed color temperature image corresponding to each white balance label image based on N kinds of preset ambient light sources to obtain N third images; the third image corresponds to a preset environment light source one by one; and training a preset deep learning model based on the N third images and the corresponding white balance label images to obtain a deep learning white balance correction model.
In specific implementation, the processor 501 and the memory 502 described in the embodiment of the present invention may execute the implementation described in the method embodiment of fig. 1 or fig. 2 provided in the embodiment of the present invention, and may also execute the implementation method of the image white balance correction apparatus described in fig. 4 provided in the embodiment of the present invention, which is not described herein again.
The embodiment of the present application further provides a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, the computer program may be used to implement the image white balance correction method described in the embodiment corresponding to fig. 1 or fig. 2 in the embodiment of the present application, and details of the method are not repeated here.
The computer readable storage medium may be an internal storage unit of the computer device of any of the foregoing embodiments, such as a hard disk or a memory of the device. The computer readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk provided on the device, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like. Further, the computer-readable storage medium may also include both an internal storage unit and an external storage device of the computer device. The computer readable storage medium is used to store computer programs and other programs and data for a computer device. The computer-readable storage medium may also be used to temporarily store data that has been output or is to be output.
It will be understood by those skilled in the art that all or part of the processes in the methods of the embodiments described above may be implemented by hardware that is instructed by a computer program, and the program may be stored in a readable storage medium, and when executed, may include the processes of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), or the like.
The above disclosure is only for the purpose of illustrating the preferred embodiments of the present application and is not to be construed as limiting the scope of the present application, so that the present application is not limited thereto, and all equivalent variations and modifications can be made to the present application.
Claims (18)
1. An image white balance correction method, characterized in that the method comprises:
acquiring an image to be calibrated;
rendering the image to be calibrated based on N kinds of preset ambient light sources to obtain N first images; the first images correspond to preset ambient light sources one to one, and N is an integer greater than 1;
and inputting the N first images into a deep learning white balance correction model to obtain a white balance correction image corresponding to the image to be calibrated.
2. The method of claim 1, wherein the image to be calibrated comprises a single color temperature image or a mixed color temperature image.
3. The method according to claim 1 or 2, wherein the inputting the N first images into a deep learning white balance correction model to obtain a white balance correction image corresponding to the image to be calibrated includes:
inputting the N first images into the deep learning white balance correction model to obtain N first weight images; the first weight image corresponds to the first image one by one;
and performing fusion processing on the N first images based on the N first weighted images to obtain a white balance correction image corresponding to the image to be calibrated.
4. The method according to any one of claims 1-3, wherein before the inputting the N first images into the deep learning white balance correction model to obtain the white balance correction image corresponding to the image to be calibrated, the method further comprises:
acquiring a white balance label image corresponding to each monochromatic temperature image in a plurality of monochromatic temperature images;
rendering each white balance label image based on M kinds of preset environment light sources to obtain a mixed color temperature image corresponding to each white balance label image; wherein M is an integer greater than 1 and less than or equal to N;
and training a preset deep learning model based on each white balance label image and the corresponding mixed color temperature image to obtain the deep learning white balance correction model.
5. The method according to claim 4, wherein the acquiring a white balance label image corresponding to each of the plurality of monochromatic temperature images comprises:
acquiring a plurality of monochromatic temperature images; wherein each monochromatic temperature image is accompanied by a color chip;
determining a light source estimation value corresponding to each monochromatic temperature image based on the color card in each monochromatic temperature image;
and carrying out white balance correction on each monochromatic temperature image based on the light source estimation value of each monochromatic temperature image to obtain a white balance label image corresponding to each monochromatic temperature image.
6. The method according to claim 4 or 5, wherein the rendering each white balance label image based on M kinds of preset ambient light sources to obtain a mixed color temperature image corresponding to each white balance label image comprises:
rendering each white balance label image based on the M kinds of preset environment light sources to obtain M second images corresponding to each white balance label image; the second images correspond to the preset ambient light sources one to one;
acquiring M second weight images corresponding to each white balance label image;
and performing fusion processing on the M second images based on the M second weighted images to obtain a mixed color temperature image corresponding to each white balance label image.
7. The method of claim 6, wherein the training a preset deep learning model based on each white balance label image and the corresponding mixed color temperature image to obtain the deep learning white balance correction model comprises:
rendering the mixed color temperature image corresponding to each white balance label image based on the N kinds of preset environment light sources to obtain N third images; the third images correspond to the preset environment light sources one to one;
and training a preset deep learning model based on the N third images and the corresponding white balance label images to obtain the deep learning white balance correction model.
8. An image white balance correction apparatus characterized by comprising:
the image acquisition unit is used for acquiring an image to be calibrated;
the image rendering unit is used for rendering the image to be calibrated based on N kinds of preset ambient light sources to obtain N first images; the first images correspond to preset ambient light sources one to one, and N is an integer greater than 1;
and the white balance correction unit is used for inputting the N first images into a deep learning white balance correction model to obtain a white balance correction image corresponding to the image to be calibrated.
9. The apparatus of claim 8, wherein the image to be calibrated comprises a single color temperature image or a mixed color temperature image.
10. The apparatus according to claim 8 or 9, wherein the white balance correction unit is further configured to:
inputting the N first images into the deep learning white balance correction model to obtain N first weight images; the first weight image corresponds to the first image one by one;
and performing fusion processing on the N first images based on the N first weighted images to obtain a white balance correction image corresponding to the image to be calibrated.
11. The apparatus according to any one of claims 8-10, wherein before said inputting said N first images into a deep learning white balance correction model, a white balance correction image corresponding to said image to be calibrated is obtained;
the image acquisition unit is also used for acquiring a white balance label image corresponding to each single-color temperature image in the multiple single-color temperature images;
the image rendering unit is further configured to render each white balance label image based on M types of preset ambient light sources to obtain a mixed color temperature image corresponding to each white balance label image; wherein M is an integer greater than 1 and less than or equal to N;
the device further comprises a training unit, wherein the training unit is used for training a preset deep learning model based on each white balance label image and the corresponding mixed color temperature image to obtain the deep learning white balance correction model.
12. The apparatus of claim 11, wherein the image acquisition unit is further configured to:
acquiring a plurality of monochromatic temperature images; wherein each monochromatic temperature image is accompanied by a color chip;
determining a light source estimation value corresponding to each monochromatic temperature image based on the color card in each monochromatic temperature image;
and carrying out white balance correction on each monochromatic temperature image based on the light source estimation value of each monochromatic temperature image to obtain a white balance label image corresponding to each monochromatic temperature image.
13. The apparatus according to claim 11 or 12, wherein the image rendering unit is further configured to:
rendering each white balance label image based on the M kinds of preset environment light sources to obtain M second images corresponding to each white balance label image; the second images correspond to the preset ambient light sources one to one;
acquiring M second weight images corresponding to each white balance label image;
and performing fusion processing on the M second images based on the M second weighted images to obtain a mixed color temperature image corresponding to each white balance label image.
14. The apparatus of claim 13,
the image rendering unit is further configured to render the mixed color temperature image corresponding to each white balance label image based on the N preset ambient light sources to obtain N third images; the third images correspond to the preset environment light sources one to one;
and the training unit is further used for training a preset deep learning model based on the N third images and the corresponding white balance label images to obtain the deep learning white balance correction model.
15. A chip, wherein the chip is configured to:
acquiring an image to be calibrated;
rendering the image to be calibrated based on N kinds of preset ambient light sources to obtain N first images; the first images correspond to preset ambient light sources one to one, and N is an integer greater than 1;
and inputting the N first images into a deep learning white balance correction model to obtain a white balance correction image corresponding to the image to be calibrated.
16. A chip module, characterized in that it comprises a chip as claimed in claim 13.
17. A computer device comprising a processor and a memory, the processor and the memory being interconnected, wherein the memory is configured to store a computer program comprising program instructions, the processor being configured to invoke the program instructions to perform the method of any of claims 1-7.
18. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program comprising program instructions that, when executed by a processor, cause the processor to perform the method according to any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210336143.4A CN114745531A (en) | 2022-03-31 | 2022-03-31 | Image white balance correction method and related device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210336143.4A CN114745531A (en) | 2022-03-31 | 2022-03-31 | Image white balance correction method and related device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114745531A true CN114745531A (en) | 2022-07-12 |
Family
ID=82278392
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210336143.4A Pending CN114745531A (en) | 2022-03-31 | 2022-03-31 | Image white balance correction method and related device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114745531A (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109361910A (en) * | 2018-10-23 | 2019-02-19 | 浙江宇视科技有限公司 | Self-adapted white balance correction method and device |
| US20200396434A1 (en) * | 2019-06-13 | 2020-12-17 | Apple Inc. | Image White Balance Processing System and Method |
| CN113411555A (en) * | 2021-06-08 | 2021-09-17 | 珠海市杰理科技股份有限公司 | White balance processing method and device, storage medium and image capturing device |
-
2022
- 2022-03-31 CN CN202210336143.4A patent/CN114745531A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109361910A (en) * | 2018-10-23 | 2019-02-19 | 浙江宇视科技有限公司 | Self-adapted white balance correction method and device |
| US20200396434A1 (en) * | 2019-06-13 | 2020-12-17 | Apple Inc. | Image White Balance Processing System and Method |
| CN113411555A (en) * | 2021-06-08 | 2021-09-17 | 珠海市杰理科技股份有限公司 | White balance processing method and device, storage medium and image capturing device |
Non-Patent Citations (1)
| Title |
|---|
| MAHMOUD AFIFI 等: ""Auto White-Balance Correction for Mixed-Illuminant Scenes"", 《COMPUTER VISION AND PATTERN RECOGNITION》, 17 September 2021 (2021-09-17), pages 1 - 5 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113170028B (en) | Method for generating image data of machine learning based imaging algorithm | |
| KR101917404B1 (en) | Colour calibration method for an image capture device | |
| CN108712639A (en) | Image color correction method, apparatus and system | |
| EP3864838B1 (en) | Meta-learning for camera adaptive color constancy | |
| CN110009712B (en) | Image-text typesetting method and related device thereof | |
| WO2020093653A1 (en) | Color adjustment method, color adjustment device, electronic device and computer-readable storage medium | |
| US11398055B2 (en) | Color evaluation device, color evaluation method, and indication object used in color evaluation method | |
| CN112689140B (en) | White balance synchronization method and device, electronic equipment and storage medium | |
| Penczek et al. | Color error in the digital camera image capture process | |
| US20160180552A1 (en) | Method And Device Of Constructing Uniform Color Space Directly From Raw Camera RGB | |
| Mukherjee et al. | Backward compatible object detection using hdr image content | |
| Bianco et al. | Error-tolerant color rendering for digital cameras | |
| Zhao et al. | Color alignment for relative color constancy via non-standard references | |
| TW202215033A (en) | Method of determining the concentration of an analyte in a sample of a bodily fluid | |
| CN114745531A (en) | Image white balance correction method and related device | |
| CN111899239A (en) | Image processing method and device | |
| CN113596427B (en) | Image white balance improving method and device, electronic equipment and storage medium | |
| CN115205168A (en) | Image processing method, device, electronic equipment, storage medium and product | |
| Kordecki et al. | A study of vignetting correction methods in camera colorimetric calibration | |
| CN117768791B (en) | Cross-camera image white balance method and system based on neural network | |
| CN111080565B (en) | Exposure fusion method, device and storage medium based on image quality change rule | |
| US20240296530A1 (en) | Temporal noise reduction architecture | |
| Techawatcharapaikul et al. | Improved Weighted Least Square Radiometric Calibration Based Noise and Outlier Rejection by Adjacent Comparagraph and Brightness Transfer Function | |
| CN118246051B (en) | Face image data rapid encryption method based on calibration factors | |
| CN116909556A (en) | Page configuration method, page configuration device, electronic equipment and storage medium |
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 |