CN109903247B - High-precision graying method for color image based on Gaussian color space correlation - Google Patents
High-precision graying method for color image based on Gaussian color space correlation Download PDFInfo
- Publication number
- CN109903247B CN109903247B CN201910134023.4A CN201910134023A CN109903247B CN 109903247 B CN109903247 B CN 109903247B CN 201910134023 A CN201910134023 A CN 201910134023A CN 109903247 B CN109903247 B CN 109903247B
- Authority
- CN
- China
- Prior art keywords
- image
- channels
- color space
- color
- correlation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Landscapes
- Facsimile Image Signal Circuits (AREA)
- Color Image Communication Systems (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a high-precision graying method of a color image based on Gaussian color space correlation, which is implemented according to the following steps: step 1: converting the color image from an RGB color space to an LMN color space, and extracting an L component as brightness information; step 2: obtaining standard deviation sigma of L, M and N channels of LMN color space L ,σ M ,σ N And correlation coefficient rho among three channels of L, M and N LM ,ρ LN ,ρ MN (ii) a And step 3: using second order linear mapping, standard deviation σ by step 2 L ,σ M ,σ N And correlation coefficient rho LM ,ρ LN ,ρ MN Acquiring chrominance information C; and 4, step 4: and (3) adding the brightness information in the step (1) and the chrominance information C in the step (3), and outputting a gray level image after normalization. The loss of contrast after the gray scale of the brightness information is compensated by using the chrominance information, the contrast of the original color image is well kept, the gray scale image can better accord with the subjective perception of human eyes, the operation speed is high, the execution is easy, and the result has robustness.
Description
Technical Field
The invention belongs to the technical field of digital image processing, and relates to a high-precision graying method of a color image based on Gaussian color space correlation.
Background
With the continuous development of image analysis and computer vision, the need for digital image processing in almost all technical fields is avoided, and the color image graying technology is more and more emphasized. Although the vast majority of images taken are color photographs, many printers still use black and white printing. To save costs, most pictures of many publications are also grayscale images. In addition, for medical images, the color images provide little information, and the subsequent calculation by directly using the gray level images can improve the operation efficiency. In these problems, the gray-scale image preprocessing is more important. Since the gray-scale image can use less data information to represent most features of the image, the gray-scale image has many applications in image preprocessing, such as edge detection, feature extraction, etc. In order to reduce the information amount of the input image or the subsequent operation amount, the color image needs to be grayed, so that the processing speed of the subsequent algorithm can be increased, and the system efficiency can be greatly improved. In addition, there are many people who prefer black and white images that look more artistic, which also derives the use of grayscale images in artistic aesthetics, such as chinese ink-and-wash rendering, black and white photography, and the like.
Graying of a color image is a dimension reduction process for converting a three-dimensional channel into one dimension, and loss of a large amount of information cannot be avoided. How to keep the original intention and the more prominent characteristics of the reproduced color as much as possible in the limited gray scale range and make the obtained gray scale image conform to the perception of human eyes is a great technical problem. The simplest straightforward graying method is to take the luminance component in color space or to weight the sum of the three components, such as the rgb2gray function in matlab. Assuming that human vision is more sensitive to green channel information, the output is the linear sum of the fixed coefficients (0.2989, 0.5870, 0.1140) of the three channels R, G, B of the color image, or the luminance channel taken from other color spaces such as CIE LAB, HSI. Although the existing methods have small computation amount and simple algorithm, the detailed information is easy to lose, and the problem that the gray level image after the gray level of the color image can not completely retain the intention of the original color image exists.
Disclosure of Invention
The invention aims to provide a high-precision graying method of a color image based on Gaussian color space correlation, which is used for solving the problem that the gray image after graying of the color image cannot completely reserve the intention of the original color image.
The technical scheme adopted by the invention is a high-precision graying method of a color image based on Gaussian color space correlation, which is implemented according to the following steps:
step 1: converting the color image from an RGB color space to an LMN color space, and extracting an L component as brightness information;
step 2: obtaining standard deviation sigma of three channels of L, M and N of LMN color space L ,σ M ,σ N And correlation coefficient rho among three channels of L, M and N LM ,ρ LN ,ρ MN ；
And 3, step 3: using second order linear mapping, standard deviation σ by step 2 L ,σ M ,σ N And correlation coefficient rho LM ,ρ LN ,ρ MN Acquiring chrominance information C;
and 4, step 4: and (4) adding the brightness information in the step (1) and the chrominance information C in the step (3), and outputting a gray level image after normalization.
The invention is also characterized in that:
in step 1, the LMN space is a gaussian color space, and the conversion between the LMN space and the RGB color space can be accomplished by linear transformation:
l is then extracted as the luminance information of the image as the main part of the grayscale image.
Standard deviation sigma of L channel in step 2 L And correlation coefficient rho of L and M channels LM Are respectively:
where N is the number of pixels contained in the L channel, μ L Is the mean value of L channels, μ M Is the mean of the M channels; the standard deviation sigma of M and N channels is calculated by the same method M ,σ N And correlation coefficients rho of L and N channels and M and N channels LN ,ρ MN 。
In the step 3, the mapping function of the chrominance information C adopts second-order linear mapping, and the square term { L ] in the space Z 2 ，M 2 ，N 2 And f, using a strategy of weighting standard deviation of each channel, so that channels with high contrast have higher weight. For coefficients of cross terms { LM, LN, MN } in space Z, correlation coefficient ρ = { ρ = between channels is used LM ,ρ LN ,ρ MN As a weight coefficient. Normalizing the square term in space Z with each channelThe product of the differences plus the product of the cross terms and the correlation coefficient between the channels yields the chrominance information C of the image:
C＝σ L *L 2 +σ M *M 2 +σ N *N 2 +ρ LM *L*M+ρ LN *L*N+ρ MN *M*N
in step 4, the mapping function g of the gray level image is the sum of the luminance information L and the chrominance information C:
g＝L+C
and finally, obtaining a final gray scale result through normalization:
wherein g is ij For each pixel point in the grayscale image, g min ,g max Representing the minimum and maximum gray values in the gray-scale image, respectively.
The method has the advantages that the loss of the contrast ratio after the brightness information is grayed is compensated by utilizing the chrominance information, the contrast ratio of the original color image is well kept, the gray image can better accord with the subjective perception of human eyes, the operation speed is high, the method is easy to execute, and the result has robustness.
Drawings
FIG. 1 is a frame diagram of an algorithm for high-precision graying of color images based on Gaussian color space correlation according to the present invention;
FIG. 2 is a flow chart of a high-precision graying method for color images based on Gaussian color space correlation according to the present invention;
FIG. 3 is a comparison graph of the algorithm used in the high-precision graying method of color image based on Gaussian color space correlation in comparison with the results of other algorithms.
Detailed Description
The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.
The invention discloses a high-precision graying method of a color image based on Gaussian color space correlation, which is implemented according to the following steps:
step 1, converting a color space and extracting brightness information:
converting the color image from an RGB color space to an LMN color space, wherein the LMN color space is a Gaussian color space, and the conversion between the LMN color space and the RGB color space can be completed through linear transformation:
the LMN color space is divided into a luminance L channel and chrominance M, N channels. L is extracted as the luminance information of the grayscale image.
Step 2, calculating the correlation among all channels in the LMN space:
the correlation among all channels in the LMN space comprises the calculation of standard deviation sigma of L, M and N channels L ,σ M ,σ N And correlation coefficient rho between channels LM ,ρ LN ,ρ MN . Standard deviation sigma of L channel L And correlation coefficient rho of L and M channels LM Are respectively:
where N is the number of pixels contained in the L channel, μ L Is the mean value of L channels, μ M Is the mean of the M channels. The standard deviation sigma of M and N channels is calculated by the same method M ,σ N And correlation coefficients rho of L and N channels and M and N channels LN ,ρ MN 。
Step 3, calculating chrominance information:
since the larger the standard deviation is, the larger the dynamic transformation range of the gray value of each channel is, the larger the contrast is, the more abundant the information contained in the channel is, and the square term { L ] in the space Z is 2 ，M 2 ，N 2 Using a strategy of weighting each standard deviation to make contrastMore heavily weighted channels have higher weights. For coefficients of cross terms { LM, LN, MN } in space Z, correlation coefficient ρ = { ρ = between channels is used LM ,ρ LN ,ρ MN As a weight coefficient, when the correlation degree of two channels is high, a higher weight is given. The product of the square term in space Z and the standard deviation of each channel plus the product of the cross term and the correlation coefficient between each channel yields the chrominance information C of the image:
C＝σ L *L 2 +σ M *M 2 +σ N *N 2 +ρ LM *L*M+ρ LN *L*N+ρ MN *M*N
step 4, outputting a gray level image:
the mapping function g of the grayscale image is the sum of luminance information L and chrominance information C:
g＝L+C
and then obtaining a final gray scale result through normalization:
wherein g is ij For each pixel point in the gray scale image, g min ,g max Representing the minimum and maximum gray values in the gray-scale image, respectively.
The present invention adopts a gaussian color model, as shown in fig. 1. A linear mapping model of three channels of brightness L, chroma M and N is established by using an LMN space of a color image. Then the brightness channel L of the model is used as the brightness information of the gray image in the first stage, which is the main part of the gray image, and the chrominance information is added in the second stage to be used as the compensation of the gray value in the first stage, so as to increase the contrast of the image. In the second stage of chroma information mapping, a second order linear mapping function is selected, and the standard deviation of the brightness component L and the chroma components M and N and the correlation coefficient between each two components are calculated and used as the coefficient of the second order mapping function. And finally, adding the first-order brightness component and the second-order chromaticity correlation to obtain a final gray result. The algorithm flow of the method of the present invention is shown in fig. 2.
Compared with the prior art, the gray level method based on the Gaussian color space correlation has the following advantages: the traditional second-order linear mapping brings difficulty to the optimization and solution of the parameters due to the increase of the parameters, the operation speed is very slow, and the phenomenon of overfitting is easy to occur, so that the graying result is unstable, and the practical use of the graying result is hindered to a great extent. The algorithm of the chapter directly uses the correlation of each channel in the LMN space as the coefficient of each element in the Z space, and has the advantages of high calculation speed, easy execution, robustness of the result and accordance with the perception of human eyes.
In order to verify the effectiveness of the method, the performance of the high-precision graying method is qualitatively and quantitatively evaluated by the following method.
FIG. 3 is a graph of the method of the present invention and the rgb2gray function in matlab, the contrast preserving algorithm (CP algorithm) proposed by Lu, the semi-parametric semi-optimization algorithm SPDecolor proposed by Liu
The results of the image set are compared to the CSDD image set. Column a is the primary color image, with 1-3 lines from
The image set, 4-8 lines from the CSDD image set. Column b is the rgb2gray result, column c is the CP algorithm proposed by Lu et al, column d is the SPDecolor algorithm proposed by Liu et al, and column e is the graying result of the algorithm proposed by the present invention. As can be seen from the figure, the rgb2gray result graph of the b column loses a great deal of contrast information, the display result is relatively flat, the gray result of the third row and the fourth row loses the structural characteristics of the original color image, but the rgb2gray result has stability. The c-column CP algorithm can well maintain the image contrast under certain conditions, such as the third row and the fourth row, but the gray scale image of the third row does not accord with human perception; and when the poster image of the 6 th row of the first row of the fruit image is too bright in brightness, the edges of the images of the fourth and seventh rows are too sharp, and the gray scale result of the eighth row is an artifact phenomenon. The d-column SPDecolor algorithm also has the situation of contrast loss, such as the second row of flowers and leaves; first, theThe SPDecolor in the third row maps the original gray background to white, and the SPDecolor in the fourth row excessively sharpens the edge and loses the original contrast; the nose tip of the sixth row of poster characters also lost the original structure. e column the algorithm provided by the invention can make the gray level image accord with the subjective perception of human eyes while keeping the contrast.
From the analysis results, the algorithm provided by the invention can better retain the image details, the contrast of the image result after graying is more obvious, the gray result has stability, and the phenomena of artifacts and the like can not occur.
In quantitative evaluation, a Cadik data set and a CSDD data set are used for respectively measuring the average ACCPR of CCPR when the values of color contrast retention CCPR and tau after graying are 1 to 15 and the value range of CCPR is [0,1]And the closer its value is to 1, the better the graying effect is represented. The results are shown in Table 1, the left and right sides of Table 1 being
CCPR values of the image set and CCPR values of the CSDD image and the CCPR values.
The image patterns contained in the image set are simple and can be seen on the left side of Table 1
Our approach in CCPR of image sets outperforms others at most threshold levels. This is consistent with subjective measurements. In the results of the CSDD images and the results on the right side of table 1, since the CSDD includes images with complex patterns and rich colors, the CCPR values of several algorithms are not greatly different, and the CCPR value of the algorithm in this chapter is closer to the SPDecolor.
TABLE 1
The running time comparison between the CP algorithm adopting the second-order linear mapping and the SPDecor algorithm is performed to prove the operation efficiency of the algorithm herein, and the result is shown in table 2. The platform used in the experiment was a 2.20GHz Corei5CPU notebook, implemented in MATLAB 2016 a. As can be seen from table 2, for images with different resolutions, the running speed of the algorithm adopted by the present invention is faster than that of other methods, and especially, as the resolution increases, the speed is improved more significantly, which ensures the real-time performance of the algorithm adopted by the present invention.
Graying algorithm runtime comparison
TABLE 2
Through the analysis of the experimental image and the experimental data, the invention is demonstrated that the image contrast can be well kept, and the grayed image is in accordance with the perception of human eyes. In addition, the method has the advantages of high running speed, difficult generation of phenomena of false shadow and the like in the gray scale result and better stability.
Claims (4)
1. The high-precision graying method of the color image based on the Gaussian color space correlation is characterized by comprising the following steps of:
step 1: converting the color image from an RGB color space to an LMN color space, and extracting an L component as brightness information;
step 2: obtaining standard deviation sigma of L, M and N channels of LMN color space L ,σ M ,σ N And correlation coefficient rho among three channels of L, M and N LM ,ρ LN ,ρ MN ；
And 3, step 3: using second order linear mapping, passing the standard deviation σ of step 2 L ,σ M ,σ N And correlation coefficient rho LM ,ρ LN ,ρ MN Obtaining chrominance information C;
and 4, step 4: adding the brightness information of the step 1 and the chrominance information C of the step 3, and outputting a gray level image after normalization;
the mapping function of the chrominance information C in the step 3 adopts second-order linear mapping, and the square term { L ] in the space Z 2 ，M 2 ，N 2 A strategy of weighting standard deviation of each channel is used, so that channels with large contrast have higher weight; for coefficients of cross terms { LM, LN, MN } in space Z, correlation coefficient ρ = { ρ = between channels is used LM ,ρ LN ,ρ MN As a weight coefficient; the chrominance information C of the image is obtained by adding the product of the square term and the standard deviation of each channel in the space Z and the product of the cross term and the correlation coefficient between each channel, and the calculation formula is as follows:
C＝σ L *L 2 +σ M *M 2 +σ N *N 2 +ρ LM *L*M+ρ LN *L*N+ρ MN *M*N。
2. the method for graying color image with high precision based on gaussian color space correlation according to claim 1, wherein the LMN space in step 1 is gaussian color space, and the conversion with RGB color space can be accomplished by the following linear transformation:
l is then extracted as the luminance information of the image as the main part of the grayscale image.
3. The color image high-precision graying method based on Gaussian color space correlation as claimed in claim 1, wherein the standard deviation σ of L channel in step 2 L And correlation coefficient rho of L and M channels LM Are respectively:
where N is the number of pixels contained in the L channel, mu L Is the mean value of L channels, μ M Is the mean of the M channels; the standard deviation sigma of M and N channels is obtained by the same method M ,σ N And correlation coefficients rho of L and N channels and M and N channels LN ,ρ MN 。
4. The method for graying color images based on Gaussian color space correlation with high precision as claimed in claim 1, wherein in the step 4, a mapping function g of the grayscale image is obtained by the sum of the luminance information L and the chrominance information C:
g＝L+C
obtaining a final gray scale result through normalization:
wherein g is ij For each pixel point in the grayscale image, g min ,g max Representing the minimum and maximum gray values in the gray-scale image, respectively.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910134023.4A CN109903247B (en) | 2019-02-22 | 2019-02-22 | High-precision graying method for color image based on Gaussian color space correlation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910134023.4A CN109903247B (en) | 2019-02-22 | 2019-02-22 | High-precision graying method for color image based on Gaussian color space correlation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109903247A CN109903247A (en) | 2019-06-18 |
| CN109903247B true CN109903247B (en) | 2023-02-03 |
Family
ID=66945503
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910134023.4A Active CN109903247B (en) | 2019-02-22 | 2019-02-22 | High-precision graying method for color image based on Gaussian color space correlation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109903247B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112801922B (en) * | 2021-04-01 | 2021-07-27 | 暨南大学 | Color image-gray image-color image conversion method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104077612A (en) * | 2014-07-15 | 2014-10-01 | 中国科学院合肥物质科学研究院 | Pest image recognition method based on multi-feature sparse representation technology |
| CN104200239A (en) * | 2014-09-09 | 2014-12-10 | 河海大学常州校区 | Image feature fusion identification based signature authentic identification system and method |
| CN104794690A (en) * | 2015-04-01 | 2015-07-22 | 国家电网公司 | Three-dimensional image color-to-gray correcting method based on SIFT (scale invariant feature transform) matching |
| CN105528609A (en) * | 2014-09-28 | 2016-04-27 | 江苏省兴泽实业发展有限公司 | Vehicle license plate location method based on character position |
| CN106503719A (en) * | 2016-09-27 | 2017-03-15 | 深圳增强现实技术有限公司 | A kind of object color is extracted and detection method and device |
| CN106780379A (en) * | 2016-12-08 | 2017-05-31 | 哈尔滨工业大学 | The microscopical colour-image reinforcing method of one kind metering |
Family Cites Families (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6504551B1 (en) * | 1997-03-14 | 2003-01-07 | Sony Corporation | Color correction device, color correction method, picture processing device, and picture processing method |
| JP2005295333A (en) * | 2004-04-01 | 2005-10-20 | Toshiba Corp | Apparatus and method for image forming, and program |
| CN101441763B (en) * | 2008-11-11 | 2012-05-16 | 浙江大学 | Multiple-colour tone image unity regulating method based on color transfer |
| CN101651772B (en) * | 2009-09-11 | 2011-03-16 | 宁波大学 | Method for extracting video interested region based on visual attention |
| CN101916370B (en) * | 2010-08-31 | 2012-04-25 | 上海交通大学 | Method for processing non-feature regional images in face detection |
| EP2671374B1 (en) * | 2011-01-31 | 2015-07-22 | Dolby Laboratories Licensing Corporation | Systems and methods for restoring color and non-color related integrity in an image |
| US9025867B2 (en) * | 2012-05-31 | 2015-05-05 | Apple Inc. | Systems and methods for YCC image processing |
| CN102723054A (en) * | 2012-06-18 | 2012-10-10 | 西安电子科技大学 | Online calibration system and online calibration method for ununiformity of LED (light-emitting diode) display screen |
| CN104182957B (en) * | 2013-05-21 | 2017-06-20 | 北大方正集团有限公司 | Traffic video information detecting method and device |
| CN103595982A (en) * | 2013-11-07 | 2014-02-19 | 天津大学 | Color image collection device based on gray level sensor and color image sensor |
| CN104181546B (en) * | 2014-08-25 | 2017-01-11 | 中国科学院武汉物理与数学研究所 | Color information acquisition and display method of color three-dimensional scanning laser radar |
| CN104361593B (en) * | 2014-11-14 | 2017-09-19 | 南京大学 | A kind of color image quality evaluation method based on HVS and quaternary number |
| CN104809710B (en) * | 2015-05-14 | 2018-02-06 | 上海兆芯集成电路有限公司 | Image de-noising method and the device using this method |
| CN106023268B (en) * | 2016-05-30 | 2019-01-11 | 南昌大学 | A kind of color image gray processing method based on the optimization of two step Parameter Subspaces |
| CN107192995B (en) * | 2017-05-23 | 2020-11-06 | 西北工业大学 | Multi-level information fusion pure orientation underwater target tracking algorithm |
| CN107392967B (en) * | 2017-07-05 | 2018-07-10 | 温州大学 | A kind of coloured image gray processing method based on multimodal gauss of distribution function |
| CN107578399B (en) * | 2017-07-25 | 2020-12-15 | 天津大学 | Full-reference image quality evaluation method based on boundary feature segmentation |
| CN108010024B (en) * | 2017-12-11 | 2021-12-07 | 宁波大学 | Blind reference tone mapping image quality evaluation method |
| CN108288258B (en) * | 2018-04-23 | 2021-08-10 | 电子科技大学 | Low-quality image enhancement method under severe weather condition |
| CN108961227B (en) * | 2018-06-22 | 2021-05-25 | 西安理工大学 | Image quality evaluation method based on multi-feature fusion of airspace and transform domain |
-
2019
- 2019-02-22 CN CN201910134023.4A patent/CN109903247B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104077612A (en) * | 2014-07-15 | 2014-10-01 | 中国科学院合肥物质科学研究院 | Pest image recognition method based on multi-feature sparse representation technology |
| CN104200239A (en) * | 2014-09-09 | 2014-12-10 | 河海大学常州校区 | Image feature fusion identification based signature authentic identification system and method |
| CN105528609A (en) * | 2014-09-28 | 2016-04-27 | 江苏省兴泽实业发展有限公司 | Vehicle license plate location method based on character position |
| CN104794690A (en) * | 2015-04-01 | 2015-07-22 | 国家电网公司 | Three-dimensional image color-to-gray correcting method based on SIFT (scale invariant feature transform) matching |
| CN106503719A (en) * | 2016-09-27 | 2017-03-15 | 深圳增强现实技术有限公司 | A kind of object color is extracted and detection method and device |
| CN106780379A (en) * | 2016-12-08 | 2017-05-31 | 哈尔滨工业大学 | The microscopical colour-image reinforcing method of one kind metering |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109903247A (en) | 2019-06-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109872285B (en) | Retinex low-illumination color image enhancement method based on variational constraint | |
| Wang et al. | Simple low-light image enhancement based on Weber–Fechner law in logarithmic space | |
| CN104537634B (en) | The method and system of raindrop influence is removed in dynamic image | |
| CN107680056B (en) | Image processing method and device | |
| CN111275696B (en) | Medical image processing method, image processing method and device | |
| CN109147005B (en) | Self-adaptive dyeing method and system for infrared image, storage medium and terminal | |
| CN106897972A (en) | A kind of self-adapting histogram underwater picture Enhancement Method of white balance and dark primary | |
| CN106780417A (en) | A kind of Enhancement Method and system of uneven illumination image | |
| CN104966285B (en) | A kind of detection method of salient region | |
| CN110298792B (en) | Low-illumination image enhancement and denoising method, system and computer equipment | |
| CN104182947A (en) | Low-illumination image enhancement method and system | |
| CN106296620B (en) | A kind of color rendition method based on histogram translation | |
| CN111968041A (en) | Self-adaptive image enhancement method | |
| CN111223110B (en) | Microscopic image enhancement method and device and computer equipment | |
| Chen et al. | Blind quality index for tone-mapped images based on luminance partition | |
| CN107256539B (en) | Image sharpening method based on local contrast | |
| CN109903247B (en) | High-precision graying method for color image based on Gaussian color space correlation | |
| CN103996179B (en) | Fast real-time image enhancement method based on single-scale Retinex | |
| CN108711160A (en) | A kind of Target Segmentation method based on HSI enhancement models | |
| CN108550124B (en) | Illumination compensation and image enhancement method based on bionic spiral | |
| CN104504666B (en) | A kind of tone mapping method based on laplacian pyramid | |
| CN113129300A (en) | Drainage pipeline defect detection method, device, equipment and medium for reducing false detection rate | |
| CN110415816B (en) | Skin disease clinical image multi-classification method based on transfer learning | |
| CN112308793A (en) | Novel method for enhancing contrast and detail of non-uniform illumination image | |
| CN116630198A (en) | Multi-scale fusion underwater image enhancement method combining self-adaptive gamma correction |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |