WO2022156129A1 - 图像处理方法、图像处理装置及计算机设备 - Google Patents

图像处理方法、图像处理装置及计算机设备 Download PDF

Info

Publication number
WO2022156129A1
WO2022156129A1 PCT/CN2021/100153 CN2021100153W WO2022156129A1 WO 2022156129 A1 WO2022156129 A1 WO 2022156129A1 CN 2021100153 W CN2021100153 W CN 2021100153W WO 2022156129 A1 WO2022156129 A1 WO 2022156129A1
Authority
WO
WIPO (PCT)
Prior art keywords
preset
picture
color
processed
data set
Prior art date
Application number
PCT/CN2021/100153
Other languages
English (en)
French (fr)
Inventor
符采灵
陈云娜
何振伟
Original Assignee
Tcl华星光电技术有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Tcl华星光电技术有限公司 filed Critical Tcl华星光电技术有限公司
Priority to US17/600,328 priority Critical patent/US20230105393A1/en
Publication of WO2022156129A1 publication Critical patent/WO2022156129A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
    • H04N1/46—Colour picture communication systems
    • H04N1/56—Processing of colour picture signals
    • H04N1/60—Colour correction or control
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/80—Geometric correction
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
    • H04N1/46—Colour picture communication systems
    • H04N1/56—Processing of colour picture signals
    • H04N1/60—Colour correction or control
    • H04N1/6077—Colour balance, e.g. colour cast correction
    • H04N1/608—Colour balance, e.g. colour cast correction within the L, C1, C2 colour signals
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/77—Retouching; Inpainting; Scratch removal
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/90—Dynamic range modification of images or parts thereof
    • G06T5/92—Dynamic range modification of images or parts thereof based on global image properties
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/90—Determination of colour characteristics
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/56—Extraction of image or video features relating to colour
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/60—Extraction of image or video features relating to illumination properties, e.g. using a reflectance or lighting model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G—PHYSICS
    • G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
    • G09G—ARRANGEMENTS OR CIRCUITS FOR CONTROL OF INDICATING DEVICES USING STATIC MEANS TO PRESENT VARIABLE INFORMATION
    • G09G3/00—Control arrangements or circuits, of interest only in connection with visual indicators other than cathode-ray tubes
    • G09G3/20—Control arrangements or circuits, of interest only in connection with visual indicators other than cathode-ray tubes for presentation of an assembly of a number of characters, e.g. a page, by composing the assembly by combination of individual elements arranged in a matrix no fixed position being assigned to or needed to be assigned to the individual characters or partial characters
    • G09G3/34—Control arrangements or circuits, of interest only in connection with visual indicators other than cathode-ray tubes for presentation of an assembly of a number of characters, e.g. a page, by composing the assembly by combination of individual elements arranged in a matrix no fixed position being assigned to or needed to be assigned to the individual characters or partial characters by control of light from an independent source
    • G09G3/36—Control arrangements or circuits, of interest only in connection with visual indicators other than cathode-ray tubes for presentation of an assembly of a number of characters, e.g. a page, by composing the assembly by combination of individual elements arranged in a matrix no fixed position being assigned to or needed to be assigned to the individual characters or partial characters by control of light from an independent source using liquid crystals
    • G—PHYSICS
    • G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
    • G09G—ARRANGEMENTS OR CIRCUITS FOR CONTROL OF INDICATING DEVICES USING STATIC MEANS TO PRESENT VARIABLE INFORMATION
    • G09G5/00—Control arrangements or circuits for visual indicators common to cathode-ray tube indicators and other visual indicators
    • G09G5/02—Control arrangements or circuits for visual indicators common to cathode-ray tube indicators and other visual indicators characterised by the way in which colour is displayed
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10024—Color image
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20076—Probabilistic image processing
    • G—PHYSICS
    • G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
    • G09G—ARRANGEMENTS OR CIRCUITS FOR CONTROL OF INDICATING DEVICES USING STATIC MEANS TO PRESENT VARIABLE INFORMATION
    • G09G2320/00—Control of display operating conditions
    • G09G2320/02—Improving the quality of display appearance
    • G09G2320/0242—Compensation of deficiencies in the appearance of colours
    • G—PHYSICS
    • G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
    • G09G—ARRANGEMENTS OR CIRCUITS FOR CONTROL OF INDICATING DEVICES USING STATIC MEANS TO PRESENT VARIABLE INFORMATION
    • G09G2320/00—Control of display operating conditions
    • G09G2320/02—Improving the quality of display appearance
    • G09G2320/028—Improving the quality of display appearance by changing the viewing angle properties, e.g. widening the viewing angle, adapting the viewing angle to the view direction
    • G—PHYSICS
    • G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
    • G09G—ARRANGEMENTS OR CIRCUITS FOR CONTROL OF INDICATING DEVICES USING STATIC MEANS TO PRESENT VARIABLE INFORMATION
    • G09G2340/00—Aspects of display data processing

Definitions

  • the present invention relates to the technical field of image processing, in particular to an image processing method, an image processing device and computer equipment.
  • Liquid crystal display technology is widely used, involving smart home, commercial display, e-sports entertainment and so on. With the development of liquid crystal display technology, the requirements for picture display quality are getting higher and higher. Therefore, it is particularly important to improve the liquid crystal display technology. In terms of image display, there are obvious differences between the hue, saturation, etc. of some colors and the real image, which will affect the quality of image display and even cause picture distortion.
  • VAC View Angle Compensation
  • the VAC algorithm can have a variety of driving modes. There are two common driving modes at present. In one of the driving modes, the picture quality display effect is better and the texture is small, but the effect of improving the color cast phenomenon is low; the other driving mode In the mode, the effect of improving the color cast is good, but the texture is large, and the picture display quality needs to be further improved.
  • Embodiments of the present invention provide an image processing method, an image processing apparatus, and a computer device, so as to solve the technical problem in the prior art that the texture and color cast cannot be simultaneously optimized to a greater extent during image processing.
  • the present invention provides an image processing method, the image processing method comprising:
  • the color output correction value of the preset color of the preset scene in the picture area is obtained, so as to perform the color output correction of the preset color in the preset scene of the picture to be processed. Compensation processing.
  • the image processing method also includes:
  • a color output correction value of the preset color of the preset scene in the picture area is obtained, so as to adjust the color output in the preset scene of the picture to be processed.
  • the preset color is subjected to compensation processing.
  • the method before acquiring the Gaussian probability of the preset color of the preset scene according to the first to-be-processed chrominance data set and the second to-be-processed chrominance data set, the method further includes:
  • the acquiring the Gaussian probability of the preset color of the preset scene according to the first to-be-processed chrominance data set and the second to-be-processed chrominance data set includes:
  • the Gaussian probability of the preset color of the preset scene is acquired by the Gaussian model.
  • obtaining the first initial chromaticity data set and the second initial chromaticity data set of the preset colors in the preset scene of the preprocessed picture including:
  • establishing a Gaussian model about the preset color according to the first initial chromaticity data set and the second initial chromaticity data set including:
  • the Gaussian model is established from the covariance matrix, the inverse of the covariance matrix, and the rank of the covariance matrix.
  • the preset scene includes one or more of portraits, blue sky, grass, food, animals and buildings.
  • the acquisition of the first initial chromaticity data set and the second initial chromaticity data set of the preset colors in the preset scene of the preprocessed picture includes:
  • the skin color data is decomposed to obtain a first initial chromaticity data set and a second initial chromaticity data set of the skin color data.
  • the obtaining the first to-be-processed chromaticity data set and the second to-be-processed chromaticity data set of the preset color in the to-be-processed picture includes:
  • acquiring the Gaussian probability of the preset color of the preset scene includes:
  • the Gaussian probability of the preset scene in the picture to be processed is acquired.
  • the acquiring the Gaussian probability of the preset color of the preset scene further includes:
  • a correlation coefficient of the preset scene in the picture to be processed is assigned a value of 0.
  • identifying the picture area containing the preset color of the preset scene in the picture to be processed includes:
  • the picture area is a picture area containing the preset color of the preset scene.
  • the obtaining the pixel gradient of any of the picture regions includes:
  • Pixel gradients of the picture region are calculated based on the pixel values.
  • acquiring the color output correction value of the preset color of the preset scene in the picture area according to the Gaussian probability including:
  • For any one of the picture regions obtain the product of the Gaussian probability of the picture region with respect to the preset color of the preset scene and the first texture probability or the second texture probability, and obtain all the pictures in the picture region.
  • the color output correction value of the preset color of the preset scene is
  • acquiring the color output correction value of the preset color of the preset scene in the picture area according to the Gaussian probability including:
  • the identifying the picture area containing the preset color of the preset scene in the picture to be processed includes:
  • Correction or suppression adjustment is performed on the brightness data of the preset color in the brightness interval to obtain a brightness adjustment probability.
  • an embodiment of the present invention further provides an image processing apparatus, including:
  • a data acquisition unit configured to acquire a first to-be-processed chromaticity data set and a second to-be-processed chromaticity data set about a preset color in the to-be-processed picture
  • a first data processing unit configured to obtain the Gaussian probability of the preset color of the preset scene from the constructed Gaussian model according to the first to-be-processed chrominance data set and the second to-be-processed chrominance data set ;
  • a texture identification unit configured to identify a picture area containing the preset scene in the picture to be processed
  • a second data processing unit configured to acquire, according to the Gaussian probability, a color output correction value of the preset color of the preset scene in the picture area, so as to adjust the color output of the preset scene of the picture to be processed.
  • the preset color in the compensation process is performed.
  • the image processing apparatus further includes a brightness adjustment unit, and the brightness adjustment unit is configured to acquire brightness data about the preset color in the picture to be processed.
  • an embodiment of the present invention also provides a computer device, including:
  • processors one or more processors
  • One or more application programs wherein the one or more application programs are stored in the memory and configured to be executed by the processor to implement the image processing method described above.
  • the computer equipment further includes a power supply and an input unit, and the power supply is logically connected to the processor through a power management system.
  • the image processing method, image processing device and computer equipment of the present invention according to whether the picture to be processed contains a preset scene, is corrected by the correlation coefficient of the corresponding preset scene, and the preset color of the preset scene obtained according to the Gaussian model is adjusted.
  • the Gaussian probability and then use the corresponding compensation processing technology, such as the use of viewing angle compensation processing, so that when the picture is displayed, the display transition of the preset color is natural, which is in line with the visual characteristics of the human eye, can effectively improve the color cast phenomenon, and can also improve the treatment process.
  • the processing efficiency of pictures improves the accuracy of color detection to avoid false detection of colors in other scenes that are similar to the preset colors in the preset scene.
  • the frame feeling can be effectively reduced and the picture quality can be improved.
  • FIG. 1 is a schematic diagram of an image processing method provided in an embodiment of the present invention.
  • Fig. 2 is the Gaussian model simulation effect diagram of the image processing method provided in the embodiment of the present invention.
  • Fig. 3 is the probability effect diagram of the skin color color processing to the portrait scene of the image processing method provided in the embodiment of the present invention
  • 4 is a probabilistic effect diagram of processing the skin color of the portrait scene, the blue color of the blue sky scene, and the green color of the grass scene by the image processing method provided in the embodiment of the present invention
  • FIG. 5 is a schematic structural diagram of a computer device in an embodiment of the present invention.
  • first and second are used for descriptive purposes only, and should not be understood as indicating or implying relative importance or implying the number of indicated technical features. Thus, features defined as “first”, “second” may expressly or implicitly include one or more features. In the description of the present invention, “plurality” means two or more, unless otherwise expressly and specifically defined.
  • the word "exemplary” is used to mean “serving as an example, illustration or illustration”. Any embodiment of this disclosure described as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments.
  • the following description is presented to enable any person skilled in the art to make and use the present invention. In the following description, details are set forth for the purpose of explanation. It will be understood that one of ordinary skill in the art will realize that the present invention may be practiced without the use of these specific details. In other instances, well-known structures and procedures have not been described in detail so as not to obscure the description of the present invention with unnecessary detail. Thus, the present invention is not intended to be limited to the embodiments shown but is to be accorded the widest scope consistent with the principles and features disclosed herein. Unless otherwise specified, the parallel or perpendicular in the orientations involved in the present invention are not parallel or perpendicular in the strict sense, as long as the corresponding structure can achieve the corresponding purpose.
  • an embodiment of the present invention provides an image processing method, and the image processing method includes:
  • Step S1 obtaining the first to-be-processed chromaticity data set and the second to-be-processed chromaticity data set about the preset color in the to-be-processed picture;
  • Step S2 obtaining the Gaussian probability of the preset color of the preset scene according to the first to-be-processed chrominance data set and the second to-be-processed chrominance data set;
  • Step S3 identifying the picture area containing the preset color of the preset scene in the picture to be processed
  • the color output correction value of the preset color of the preset scene in the picture area is obtained, so as to perform compensation processing for the preset color in the preset scene of the picture to be processed.
  • the constructed Gaussian model processes the relevant data in the picture to be processed to obtain the Gaussian probability of the preset color of the preset scene in the picture to be processed.
  • a plurality of preprocessed pictures can be selected from the relevant database, and these preprocessed pictures respectively contain corresponding preset scenes.
  • the type and number of preset scenes can be specifically set according to actual needs.
  • the preset scenes can be portraits, blue sky, grass, food, animals, other natural scenery, buildings.
  • the corresponding preset color is the corresponding color in each scene. For example, in a portrait scene, the preset color can be skin color; in a blue sky scene, the preset color can be blue; in a grass scene, the preset color can be green.
  • the sensitive color of the human eye and the corresponding scene are selected as examples for description.
  • three preset scenes of portrait, blue sky and grass are selected, and the corresponding preset colors are skin color, blue and green respectively.
  • the following three preset scenes and corresponding preset colors are used as examples for description.
  • a plurality of first preprocessed pictures containing a preset portrait scene, a plurality of second preprocessed pictures containing a blue sky preset scene, and a plurality of third preprocessed pictures containing a grass preset scene are respectively selected from the database.
  • the number of preprocessed pictures containing each preset scene is specifically set according to the actual situation.
  • the skin color data of the first preprocessed picture is extracted, and the specific extraction method may adopt the current conventional extraction method.
  • the skin color data can be decomposed in the Ycbcr space to obtain luminance data, first initial chromaticity data and second initial chromaticity data about the skin color data, respectively.
  • a luminance data set, a first initial chromaticity data set and a second initial chromaticity data set related to the skin color data can be obtained.
  • the decomposition processing of the skin color data can be processed in the Ycbcr space, or in the HSB color space, etc.; similarly, the following methods of decomposing the preset colors of other preset scenes can be processed in the Ycbcr space. It can also be processed in other color spaces. The following is an example of processing in the Ycbcr space.
  • the skin color data of the first preprocessed picture can be processed by the following formula:
  • R, G, B in the above formula are the red component value, green component value and blue component value of the skin color data, respectively, y skin(i) is the brightness data of the skin color data, cb skin(i) is the skin color data.
  • the first initial chromaticity data, cr skin(i) is the second initial chromaticity data of the skin color data.
  • Obtain the mean value of the first initial chromaticity data set obtain the first chromaticity mean value ⁇ skin1 about skin color data; and obtain each first initial chromaticity data in the first initial chromaticity data set and the above-mentioned first chromaticity mean value The variance a skin between ⁇ skin1 .
  • Obtain the mean value of the second initial chromaticity data set obtain the second chromaticity mean value ⁇ skin2 about the skin color data; and obtain each second initial chromaticity data in the second initial chromaticity data set and the above-mentioned second chromaticity mean value The variance d skin between ⁇ skin2 .
  • the second initial chromaticity Data sets cr skin(1) , cr skin(2) ??cr skin(i) «, to obtain the first initial chromaticity data and the second information about the skin color in the portrait scene
  • the covariance matrix cov(cb skin ,cr skin ) of the initial chromaticity data is expressed as follows:
  • cb skin(i) is the first initial chromaticity data of any first preprocessed picture
  • crskin (i) is the second initial chromaticity data of any first preprocessed picture
  • ⁇ skin1 is a plurality of first preprocessed pictures.
  • the first chromaticity mean value of the skin color data of a preprocessed picture ⁇ skin2 is the second chromaticity mean value of the skin color data of a plurality of first preprocessed pictures
  • a skin is the first initial color of the skin color data of the first preprocessed picture
  • the variance matrix between the chromaticity data and the above-mentioned first chromaticity mean ⁇ skin1 , d skin is the variance matrix between the second initial chromaticity data of the skin color data of the first preprocessed picture and the above-mentioned second chromaticity mean ⁇ skin2
  • b skin and c skin are the correlations between the first initial chromaticity data set and the second initial chromaticity data set regarding the skin color of the first preset picture.
  • ⁇ skin can be obtained from the above formula (4).
  • a Gaussian model about the skin color in the portrait scene can be constructed, which can be specifically expressed as follows:
  • A is the amplitude of the Gaussian model
  • the value range is [0, 1]
  • gauss skin (cb i , cr i ) is the initial probability of the skin color obtained by the Gaussian model in the portrait scene
  • a skin is the first prediction.
  • the variance matrix between the first initial chromaticity data of the skin color data of the processing picture and the above-mentioned first chromaticity mean ⁇ skin1 , d skin is the second initial chromaticity data of the skin color data of the first preprocessed picture and the above-mentioned second color
  • the variance matrix between the degree mean ⁇ skin2 , cb i is the first chromaticity variable related to skin color, cr i is the second chromatic variable related to skin color, ⁇ skin -1 is cov(cb skin ,cr skin )
  • is the rank of cov(cb skin ,cr skin ), and ⁇ skin is the mean value of the first initial chromaticity data set and the second initial chromaticity data set about the skin color of the first preset picture.
  • the blue data of the second preprocessed picture is extracted.
  • the blue data can be decomposed in the Ycbcr space to obtain luminance data, first initial chrominance data and second initial chrominance data about the blue data respectively .
  • a luminance data set, a first initial chrominance data set and a second initial chrominance data set for the blue data can be obtained.
  • the blue data of the second preprocessed picture can be processed by the following formula:
  • cr sky(i) (R*0.4392+G*0.3678+B*0.0714)+128 (8)
  • R, G, B in the above formula are the red component value, green component value and blue component value of blue data respectively
  • y sky(i) is the brightness data of blue data
  • cb sky(i) is blue data
  • cr sky(i) is the second initial chromaticity data of the blue data.
  • the covariance matrix cov(cb sky ,cr sky ) of the first initial chromaticity data and the second initial chromaticity data about the blue color in the blue sky scene can be obtained from the above-mentioned data of the second preprocessed picture, and the specific expression is as follows:
  • cb sky(i) is the first initial chromaticity data of any second preprocessed picture
  • cr sky(i) is the second initial color of any second preprocessed picture.
  • Intensity data ⁇ sky1 is the first chromaticity mean value of the blue data of the plurality of second preprocessed pictures
  • ⁇ sky2 is the second chromaticity mean value of the blue data of the plurality of second preprocessed pictures
  • asky is the second preprocessing image.
  • the variance matrix between the degree mean ⁇ sky2, bsky and csky are the correlation between the first initial chrominance data set and the second initial chrominance data set;
  • A is the amplitude of the Gaussian model, the value range is [0, 1 ]
  • gausssky(cbi,cri) is the initial probability of blue color obtained by the Gaussian model in the blue sky scene, asky is the first initial chromaticity data of the blue data of the second preprocessed picture and the first chromaticity mean ⁇ sky1
  • the variance matrix between, dsky is the variance matrix between the second initial chromaticity data of the blue data of the second preprocessed picture and the above-mentioned second
  • any third preprocessed picture extract the green data of the third preprocessed picture.
  • the green data can be decomposed in the Ycbcr space to obtain luminance data, first initial chrominance data and second initial chrominance data about the green data, respectively.
  • a luminance data set, a first initial chrominance data set and a second initial chrominance data set for the green data can be obtained.
  • the green data of the third preprocessed picture can be specifically processed by the following formula:
  • cr grass(i) (R*0.4392+G*0.3678+B*0.0714)+128 (13)
  • R, G, B in the above formula are the red component value, green component value and blue component value of green data respectively
  • y grass(i) is the brightness data of green data
  • cb grass(i) is the green data
  • cr grass(i) is the second initial chromaticity data of the green data.
  • the covariance matrix cov(cb sky ,cr sky ) of the first initial chromaticity data and the second initial chromaticity data about the blue color in the blue sky scene can be obtained from the above-mentioned data of the second preprocessed picture, and the specific expression is as follows:
  • cb grass(i) is the first initial chromaticity data of any third preprocessed picture
  • cr grass(i) is the second initial color of any third preprocessed picture.
  • degree data ⁇ grass1 is the first chromaticity mean value of the green data of the plurality of third preprocessed pictures
  • ⁇ grass2 is the second chromaticity mean value of the green data of the plurality of third preprocessed pictures
  • a grass is the third preprocessed image
  • d grass is the second initial chromaticity data of the green data of the third preprocessed picture and the above-mentioned second chromaticity
  • the variance matrix between the mean ⁇ grass2 , b grass and c grass are the correlation between the first initial chromaticity data set and the second initial chromaticity data set;
  • A is the magnitude of the Gaus
  • the type and quantity of preset scenes can be set according to the requirements, and the type and quantity of preset colors can be set accordingly, and the Gaussian model of each preset color in each preset scene can be established separately.
  • the specific composition of the Gaussian model can be flexibly adjusted according to application scenarios, different customer needs or image quality requirements, etc.
  • parameters such as the amplitude value in the Gaussian model, the mean value of the preset colors in the preprocessed image, and the relevant covariance matrix can be adjusted according to the requirements, and the parameters can also be adjusted according to the accuracy or other considerations, which has strong practicality and versatility.
  • the data of the preset color of the picture to be processed is first extracted.
  • the color data of skin color, blue color and grass in the picture to be processed are extracted respectively, and decomposed in Ycbcr space respectively to obtain the color data of skin color.
  • First and second pending chrominance data sets, first and second pending chrominance data sets for blue tint, and first pending chrominance data set for green tint A chroma data set and a second pending chroma data set.
  • the first chromaticity data set to be processed and the second chromaticity data set to be processed for each preset color can be collected.
  • Each chromaticity data of is substituted into the Gaussian model of the corresponding preset color to obtain an initial probability map about the preset color.
  • the initial probability of the skin color in the picture to be processed can be obtained picture.
  • each chromaticity data in the first chromaticity data set to be processed and the second chromaticity data set to be processed corresponding to the blue color are substituted into the above formula (10), and the blue color in the picture to be processed can be obtained.
  • the chromaticity data in the first chromaticity data set to be processed and the chromaticity data in the second chromaticity data set to be processed correspondingly are substituted into the above formula (15) to obtain the initial probability map of green color in the picture to be processed.
  • the picture to be processed when the picture to be processed is processed, it may be judged whether the picture to be processed contains a preset scene, and a correlation coefficient of the preset scene may be assigned according to the judgment result. According to whether the picture to be processed contains a preset scene, the correlation coefficient of the corresponding preset scene is corrected, and the Gaussian probability of the preset color of the preset scene obtained according to the Gaussian model can be adjusted, which can not only improve the processing efficiency of the picture to be processed , and can also improve the accuracy of color detection to avoid false detection of colors in other scenes that are similar to the preset colors of the preset scene.
  • the method for judging whether the picture to be processed contains a preset scene may be processed in a conventional manner at present.
  • the comprehensive Gaussian probability value of the preset colors in the multiple preset scenes can be obtained from the initial probability corrected by the correlation coefficient of each preset scene.
  • the sum can be obtained by the following formula:
  • gauss(cb,cr) ⁇ *gauss skin (cb i ,cr i )+ ⁇ *gauss sky (cb i ,cr i )+ ⁇ *gauss grass (cb i ,cr i ) (16)
  • gauss(cb, cr) is the Gaussian probability of the preset color of the preset scene in the picture to be processed
  • ⁇ is the correlation coefficient of the portrait scene in the picture to be processed
  • gauss skin (cb i , cr i ) is the Gaussian model
  • ⁇ is the correlation coefficient of the blue sky scene in the picture to be processed
  • gauss sky (cb i , cr i ) is the initial probability of the blue color obtained by the Gaussian model
  • ⁇ is the picture to be processed
  • gauss grass (cb i ,cr i ) is the initial probability of green color obtained by the Gaussian model.
  • the corresponding correlation coefficient can be assigned as 0, then the product of it and the initial probability of the preset color of the preset scene obtained by fitting the Gaussian model is: 0 to avoid false detection of similar colors in the image to be processed.
  • the correlation coefficient ⁇ about the portrait scene is assigned a value of 1; when there is no portrait scene in the picture to be processed, the correlation coefficient ⁇ about the portrait scene is assigned a value of 0.
  • the correlation coefficient ⁇ with respect to the blue sky scene is assigned a value of 1; when there is no blue sky scene in the picture to be processed, the correlation coefficient ⁇ with respect to the blue sky scene is assigned a value of 0.
  • the correlation coefficient ⁇ about the grass scene is assigned a value of 1; when there is no grass scene in the picture to be processed, the correlation coefficient ⁇ about the grass scene is assigned a value of 0.
  • the color data of the picture to be processed is fitted by a Gaussian model and corrected by the corresponding preset scene correlation coefficient.
  • the picture to be processed contains a portrait scene, a blue sky scene and/or a grass scene, there are usually other scenes, such as the item cabinet behind the portrait in FIG. 3(a).
  • the image region in the to-be-processed picture containing the preset color of the preset scene can also be identified, so as to more accurately simulate the preset color of the preset scene in the to-be-processed picture.
  • the Gaussian probability obtained by the Gaussian model is applied, that is, the color output correction value of the preset color of the preset scene in the picture to be processed can be obtained.
  • the color output correction value is applied to the viewing angle compensation technology, which can effectively improve the color shift phenomenon of the preset color of the preset scene in the output picture.
  • the color data after fitting and processing by the Gaussian model has a natural color transition, a small sense of grid, and a higher image quality, which conforms to the visual characteristics of the human eye and provides a better viewing experience.
  • the picture to be processed when identifying the picture area of the preset scene in the picture to be processed, can be divided into multiple picture areas first, and the size of the picture area can be specifically adjusted according to the actual situation or demand, for example, the picture to be processed can be divided into Divide into 10*10 size areas.
  • any picture area extract the pixel value in the picture area to obtain the pixel gradient in the picture area; for example, current conventional algorithms such as the sobel operator can be used to calculate the pixel gradient T of the picture area.
  • a first preset threshold T thresh is set, and the pixel gradient value at each pixel is compared with the first preset threshold T thresh . For a pixel gradient value higher than the first preset threshold T thresh , the pixels corresponding to the pixel gradient value are counted.
  • the first preset threshold T thresh and the second preset threshold L thresh can be specifically set according to actual conditions or requirements.
  • each picture area of the picture to be processed is identified, and assigned values according to the identified situation.
  • the color output correction value for any picture area, the product of the corresponding Gaussian probability and the texture probability of the picture area , to obtain the color output correction value of the picture area with respect to the preset color of the preset scene. Then when the picture area contains the preset scene preset color, the color output correction value of the corresponding color in the picture area is not 0, and the color output correction value is used for the viewing angle compensation technology to improve the color cast; when the picture area does not contain When the preset color of the scene is preset, the color output correction value corresponding to the picture area is 0, and the color of the picture area is not processed.
  • the above processing method can further identify the specific area where the preset scene such as the portrait scene is located on the basis of the preliminary judgment on the preset scene such as the portrait scene. Avoid false detection of colors similar to the preset colors in the preset scene, effectively improve the accuracy of data extraction such as preset colors in the preset scene, and reduce the over-detection rate and false-detection rate.
  • the texture probability P corresponding to the picture area is assigned as 1. Moreover, only for the picture area containing the preset color of the preset scene, the corresponding color data is extracted, and the Gaussian model is used for fitting processing to obtain the Gaussian probability corresponding to the picture area, and the Gaussian probability is the color corresponding to the picture area. Output correction value.
  • the color output correction value can be implemented by the following algorithm:
  • gauss out is the color output correction value of the preset color of the preset scene in the picture to be processed
  • gauss(cb, cr) is the Gaussian probability of the preset color of the preset scene in the picture to be processed
  • P is the color to be processed.
  • the accuracy of detecting the preset color of the preset scene of the picture to be processed can be further improved according to the brightness data of the preset color of the preset scene of the picture to be processed.
  • the brightness of the preset color of the preset scene can be divided into a plurality of brightness intervals according to the brightness data, and in different brightness intervals, different linear adjustment models are used to respectively perform the brightness data of the color data of the corresponding preset color. Correcting or suppressing the adjustment can reduce the flickering problem that may occur in the data collection of low-gray-scale pictures, and can further improve the accuracy of color data processing for the preset color of the preset scene. See Figure 3 (a) for the original image, (b) of Fig. 3 is a probabilistic effect diagram of (a) using the image processing method of the embodiment of the present invention to detect the skin color of a portrait scene, and (c) of Fig. 3 is a detail diagram of Fig.
  • FIG. 4(a) is the original image
  • Fig. 4(b) is the image processing method for the skin color of the portrait scene, the blue color of the blue sky scene and the grass scene using the image processing method of the embodiment of the present invention to the picture (a). Probabilistic rendering of green color detection processing.
  • linear adjustment model can be specifically expressed as follows:
  • K(y) is the brightness adjustment probability of the brightness data of the preset color of the preset scene after being corrected by the linear adjustment model
  • y is the brightness data of the preset color of the preset scene
  • k 1 , k 2 , k 3 , k 4 , l 1 , l 2 , l 3 , and l 4 are setting parameters, respectively.
  • k 1 , k 2 , k 3 , k 4 , l 1 , l 2 , l 3 , and l 4 can be set according to the probability of setting different brightness intervals; Adjust the actual brightness of the panel, etc.
  • k 1 , k 2 , k 3 , k 4 , l 1 , l 2 , l 3 , and l 4 can be initially set according to the following assignments, respectively, and the linear adjustment model can be expressed as:
  • the obtained brightness adjustment probability and the Gaussian probability of the preset color of the preset scene in the image to be processed, and the image area of the preset color of the preset scene in the image to be processed It can effectively improve the accuracy of data processing and reduce the over-detection and false-detection rates by comprehensively processing the image to be processed, which can be realized by the following algorithm:
  • gauss out is the color output correction value of the preset color of the preset scene in the picture to be processed
  • gauss(cb, cr) is the Gaussian probability of the preset color of the preset scene in the picture to be processed
  • P is the color of the preset scene in the picture to be processed
  • K(y) is the brightness data about the preset scene preset color The brightness adjustment probability after correction by the linear adjustment model.
  • the color output correction value can be applied to the viewing angle compensation technology.
  • the specific implementation process is not specifically limited in this embodiment of the present invention.
  • the processing process thereof It can be briefly described as follows:
  • V table is the driving voltage set by the built-in table of the driving mode in the index viewing angle compensation technology
  • I RGB is the RGB value of the picture to be processed
  • LR , HR , LG , LB , and HB are the viewing angles, respectively The setting value of the RGB channel in the built-in table of drive modes in compensation technology.
  • An embodiment of the present invention also provides an image processing apparatus, including:
  • a data acquisition unit configured to acquire a first to-be-processed chromaticity data set and a second to-be-processed chromaticity data set about a preset color in the to-be-processed picture
  • a first data processing unit configured to obtain the Gaussian probability of the preset color of the preset scene by the constructed Gaussian model according to the first to-be-processed chrominance data set and the second to-be-processed chrominance data set;
  • a texture identification unit used to identify the picture area that contains the preset scene in the picture to be processed
  • the second data processing unit is configured to obtain the color output correction value of the preset color of the preset scene in the picture area according to the Gaussian probability, so as to perform compensation processing for the preset color in the preset scene of the picture to be processed.
  • a Gaussian model about the preset color is established according to the first initial chromaticity data set and the second initial chromaticity data set of the preset color in the preset scene of the preprocessed picture.
  • the image processing device further includes a brightness adjustment unit, which is used for acquiring brightness data about preset colors in the picture to be processed; judging the brightness interval in which the brightness data is located; fetching a linear adjustment model of the brightness interval; Brightness data, obtains the brightness adjustment probability of the brightness in the picture to be processed.
  • the data processing unit is configured to obtain the color output correction value of the preset color of the preset scene in the picture area according to the brightness adjustment probability and the Gaussian probability, so as to perform compensation processing for the preset color in the preset scene of the picture to be processed .
  • the image processing device uses the constructed Gaussian model to perform fitting processing on the preset color of the preset scene, and according to whether the picture to be processed contains the preset scene, corrects it by the correlation coefficient of the corresponding preset scene, and adjusts according to The Gaussian probability of the preset color of the preset scene obtained by the Gaussian model, after the image is subjected to viewing angle compensation processing, the preset color display transition is natural when the image is displayed, which is in line with the visual characteristics of the human eye, and can effectively improve the color cast phenomenon. It can also improve the processing efficiency of the picture to be processed, and improve the accuracy of color detection, so as to avoid false detection of the color similar to the preset color of the preset scene in other scenes.
  • the color data after fitting and processing by the Gaussian model has a natural color transition, a small sense of grid, and a higher image quality, which conforms to the visual characteristics of the human eye and provides a better viewing experience.
  • An embodiment of the present invention also provides a computer device, including:
  • processors one or more processors
  • One or more application programs wherein the one or more application programs are stored in the memory and configured to be executed by the processor to implement the image processing method described above.
  • the computer device may be an independent server, or may be a server network or server cluster composed of servers.
  • the computer device described in the embodiments of the present invention includes but is not limited to a computer, a network host, a single network server, a plurality of A network server set or a cloud server composed of multiple servers.
  • the cloud server is composed of a large number of computers or network servers based on cloud computing (Cloud Computing).
  • the computer device used in the embodiments of the present invention may be a device including both receiving and transmitting hardware, that is, a device having receiving and transmitting hardware capable of performing bidirectional communication on a bidirectional communication link.
  • Such devices may include cellular or other communication devices with a single-line display or a multi-line display or a cellular or other communication device without a multi-line display.
  • the specific computer equipment may specifically be a desktop terminal or a mobile terminal, and the computer equipment may specifically be one of a mobile phone, a tablet computer, a notebook computer, and the like.
  • the computer device may include a processor 2 with one or more processing cores, one or more memories 3 , a power supply 1 and an input unit 4 and other components.
  • a processor 2 with one or more processing cores, one or more memories 3 , a power supply 1 and an input unit 4 and other components.
  • FIG. 5 does not constitute a limitation on the computer device, and may include more or less components than the one shown, or combine some components, or arrange different components. in:
  • the processor 2 is the control center of the computer equipment, using various interfaces and lines to connect various parts of the entire computer equipment, by running or executing the target files and/or modules stored in the memory 3, and calling the stored in the memory 3.
  • Object files perform various functions of computer equipment and process data, so as to conduct overall monitoring of computer equipment.
  • the processor 2 may include one or more processing cores; preferably, the processor 2 may integrate an application processor and a modulation and demodulation processor, wherein the application processor mainly processes the operating system, user interface and application programs, etc. , the modem processor mainly deals with wireless communication. It can be understood that, the above-mentioned modulation and demodulation processor may also not be integrated into the processor 2 .
  • the memory 3 can be used to store target files and modules such as software programs, and the processor 2 executes various functional applications and data processing by running the target files and modules stored in the memory 3 .
  • the memory 3 may mainly include a stored program area and a stored data area, wherein the stored program area can store an operating system, an application program (such as a sound playback function, an image playback function, etc.) required for at least one function, and the like; Data created by the use of computer equipment, etc.
  • the memory 3 may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid-state storage device.
  • the memory 3 may also include a memory controller to provide access to the memory 3 by the processor 2 .
  • the computer equipment also includes a power supply 1 for supplying power to various components.
  • the power supply 1 can be logically connected to the processor 2 through a power management system, so that functions such as charging, discharging, and power consumption management are implemented through the power management system.
  • the power source 1 may also include one or more DC or AC power sources, recharging systems, power failure detection circuits, power converters or inverters, power status indicators, and any other components.
  • the computer device may also include an input unit 4 that may be used to receive input numerical or character information and generate keyboard, mouse, joystick, optical or trackball signal input related to user settings and function control.
  • an input unit 4 may be used to receive input numerical or character information and generate keyboard, mouse, joystick, optical or trackball signal input related to user settings and function control.
  • the computer device may also include a display unit and the like, which will not be described herein again.
  • the processor 2 in the computer device will load the executable files corresponding to the processes of one or more application programs into the memory 3 according to the following instructions, and the processor 2 will run them and store them in the memory 3 .
  • the image processing method uses the constructed Gaussian model to perform fitting processing on the preset color of the preset scene, and according to whether the image to be processed contains the preset scene, the corresponding The correlation coefficient of the preset scene is corrected, and the Gaussian probability of the preset color of the preset scene obtained according to the Gaussian model is adjusted.
  • the display transition of the preset color is natural when the picture is displayed. In line with the visual characteristics of the human eye, it can effectively improve the color cast phenomenon, improve the processing efficiency of the image to be processed, and improve the accuracy of color detection, so as to avoid false detection of colors similar to the preset colors in other scenes. .
  • the frame feeling can be effectively reduced and the picture quality can be improved.
  • the above units or structures can be implemented as independent entities, or can be arbitrarily combined to be implemented as the same or several entities.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computer Hardware Design (AREA)
  • Signal Processing (AREA)
  • Software Systems (AREA)
  • Crystallography & Structural Chemistry (AREA)
  • Chemical & Material Sciences (AREA)
  • Human Computer Interaction (AREA)
  • Image Processing (AREA)
  • Image Analysis (AREA)

Abstract

本发明公开了一种图像处理方法、图像处理装置及计算机设备，图像处理方法包括：获取待处理图片中预设色彩的第一待处理色度数据集和第二待处理色度数据集；获取预设色彩的高斯概率；识别待处理图片中含有预设场景的预设色彩的图片区域；获取色彩输出修正值。本发明的方法能够有效改善色偏现象，图片显示质量更佳。

Description

图像处理方法、图像处理装置及计算机设备 技术领域
本发明涉及图像处理技术领域，具体涉及一种图像处理方法、图像处理装置及计算机设备。
背景技术
液晶显示技术应用广泛，涉及智能家居、商业显示、电竞娱乐等方面。随着液晶显示技术的发展，对画面显示质量的要求也越来越高，因此，提升液晶显示技术尤其重要。在图像显示方面，某些颜色的色相、饱和度等与真实的图像之间有明显的区别，会影响图像显示的质量，甚至会造成图片的失真。
在改善色偏现象的图像处理技术中，其中常用的一种，是用于液晶显示的视角补偿(View Angle Compensation，VAC)技术，该技术能修正侧视角下的gamma曲线，解决大视角色偏问题。
VAC算法可有多种驱动模式，目前常见的两种驱动模式下，其中一种驱动模式下，图片质量显示效果较好，格感小，但是，改善色偏现象的效果低；另一种驱动模式下，色偏现象改善效果良好，但是，格感大，图片显示质量有待进一步提升。
技术问题
本发明实施例提供一种图像处理方法、图像处理装置及计算机设备，以解决现有技术中图像处理时格感与色偏现象无法同时更大程度优化的技术问题。
技术解决方案
第一方面，本发明提供一种图像处理方法，所述图像处理方法包括：
获取待处理图片中关于预设色彩的第一待处理色度数据集和第二待处理色度数据集；
根据所述第一待处理色度数据集和所述第二待处理色度数据集，获取所述预设场景的所述预设色彩的高斯概率；
识别所述待处理图片中含有所述预设场景的所述预设色彩的图片区域；
根据所述高斯概率，获取所述图片区域中所述预设场景的所述预设色彩的色彩输出修正值，以对所述待处理图片的所述预设场景中的所述预设色彩进行补偿处理。
进一步地，所述图像处理方法还包括：
获取所述待处理图片中关于所述预设色彩的亮度数据；
判断所述亮度数据所在的亮度区间；
调取所述亮度区间的线性调节模型；
根据所述线性调节模型和所述亮度数据，获取所述待处理图片中关于亮度的亮度调整概率；
根据所述亮度调整概率和所述高斯概率，获取所述图片区域中所述预设场景的所述预设色彩的色彩输出修正值，以对所述待处理图片的所述预设场景中的所述预设色彩进行补偿处理。
进一步地，在所述根据所述第一待处理色度数据集和所述第二待处理色度数据集，获取所述预设场景的预设色彩的高斯概率之前，所述方法还包括：
获取预处理图片的预设场景中预设色彩的第一初始色度数据集和第二初始色度数据集；
根据所述第一初始色度数据集和所述第二初始色度数据集，建立关于所述预设色彩的高斯模型；
所述根据所述第一待处理色度数据集和所述第二待处理色度数据集，获取所述预设场景的预设色彩的高斯概率，包括：
根据所述第一待处理色度数据集和所述第二待处理色度数据集，由所述高斯模型获取所述预设场景的所述预设色彩的高斯概率。
进一步地，获取预处理图片的预设场景中预设色彩的第一初始色度数据集和第二初始色度数据集，包括：
获取多个含有所述预设场景的预处理图片；
提取任一所述预处理图片中关于所述预设场景的所述预设色彩的色彩数据，获取所述第一初始色度数据集和所述第二初始色度数据集。
进一步地，根据所述第一初始色度数据集和所述第二初始色度数据集，建立关于所述预设色彩的高斯模型，包括：
分别获取所述第一初始色度数据集和所述第二初始色度数据集的均值；
获取关于所述第一初始色度数据集和所述第二初始色度数据集的协方差矩阵、协方差矩阵的逆以及协方差矩阵的秩；
根据所述协方差矩阵、所述协方差矩阵的逆以及所述协方差矩阵的秩，建立所述高斯模型。
进一步地，所述预设场景包括人像、蓝天、草地、食物、动物和建筑物中的一种或多种。
进一步地，所述获取预处理图片的预设场景中预设色彩的第一初始色度数据集和第二初始色度数据集，包括：
获取多个含有人像预设场景的第一预处理图片；
提取多个所述第一预处理图片的肤色数据；
对所述肤色数据进行分解处理，得到所述肤色数据的第一初始色度数据集和第二初始色度数据集。
进一步地，所述获取待处理图片中关于预设色彩的第一待处理色度数据集和第二待处理色度数据集，包括：
提取待处理图片的预设色彩的数据；
对所述预设色彩的数据进行分解处理；
获取所述预设色彩的第一待处理色度数据集和第二待处理色度数据集。
进一步地，获取所述预设场景的所述预设色彩的高斯概率，包括：
判断所述待处理图片中是否含有所述预设场景；
根据所述待处理图片中含有所述预设场景的判断结果，对所述待处理图片中关于所述预设场景的相关性系数赋值；
对于所述待处理图片中的任一所述预设色彩，根据所述高斯模型获取所述预设色彩的初始概率；
根据所述初始概率以及所述相关性系数，获取所述待处理图片中所述预设场景的高斯概率。
进一步地，所述预设场景的数量有多个，所述预设色彩有多个；
对于任一所述预设场景，所述预设场景的初始概率经相关性系数修正后，
求取所述待处理图片中多个所述预设场景的经所述相关性系数修正后的所述初始概率之和，获取所述待处理图片中关于多个所述预设场景的高斯概率。
进一步地，所述获取所述预设场景的所述预设色彩的高斯概率，还包括：
判断所述待处理图片中是否含有所述预设场景；
若所述待处理图片中没有所述预设场景，对所述待处理图片中关于所述预设场景的相关性系数赋值为0。
进一步地，识别所述待处理图片中含有所述预设场景的所述预设色彩的图片区域，包括：
将所述待处理图片划分为多个所述图片区域；
获取任一所述图片区域的像素梯度；
统计任一所述图片区域中像素梯度高于第一预设阈值的像素数量；
判断所述像素数量是否大于第二预设阈值；
若所述像素数量大于第二预设阈值，所述图片区域为含有所述预设场景所述预设色彩的图片区域。
进一步地，所述获取任一所述图片区域的像素梯度，包括：
提取任一所述图片区域内的像素值；
根据所述像素值计算所述图片区域的像素梯度。
进一步地，根据所述高斯概率，获取所述图片区域中所述预设场景的所述预设色彩的色彩输出修正值，包括：
若所述像素数量大于第二预设阈值，获取所述图片区域的第一纹理概率；
若所述像素数量小于第二预设阈值，获取所述图片区域的第二纹理概率；
对于任一所述图片区域，获取所述图片区域关于所述预设场景的所述预设色彩的高斯概率与所述第一纹理概率或第二纹理概率之积，得到所述图片区域中所述预设场景的所述预设色彩的色彩输出修正值。
进一步地，根据所述高斯概率，获取所述图片区域中所述预设场景的所述预设色彩的色彩输出修正值，包括：
获取含有所述预设场景的所述预设色彩的图片区域的第一纹理概率；
获取所述第一纹理概率与含有所述预设场景的图片区域的所述高斯概率之积，得到所述图片区域中所述预设场景的所述预设色彩的色彩输出修正值。
进一步地，所述识别所述待处理图片中含有所述预设场景的所述预设色彩的图片区域，包括：
获取所述待处理图片的亮度数据；
根据所述亮度数据将所述预设场景的所述预设色彩的亮度划分为多个亮度区间；
对所述亮度区间内的所述预设色彩的亮度数据进行修正或抑制调节，得到亮度调整概率。
第二方面，本发明实施例还提供一种图像处理装置，包括：
数据采集单元，用于获取待处理图片中关于预设色彩的第一待处理色度数据集和第二待处理色度数据集；
第一数据处理单元，用于根据所述第一待处理色度数据集和所述 第二待处理色度数据集，由所构建的高斯模型获取预设场景的所述预设色彩的高斯概率；
纹理识别单元，用于识别所述待处理图片中含有所述预设场景的图片区域；
第二数据处理单元，用于根据所述高斯概率，获取所述图片区域中所述预设场景的所述预设色彩的色彩输出修正值，以对所述待处理图片的所述预设场景中的所述预设色彩进行补偿处理。
进一步地，所述图像处理装置还包括亮度调节单元，所述亮度调节单元用于获取所述待处理图片中关于所述预设色彩的亮度数据。
第三方面，本发明实施例还提供一种计算机设备，包括：
一个或多个处理器；
存储器；以及
一个或多个应用程序，其中所述一个或多个应用程序被存储于所述存储器中，并配置为由所述处理器执行以实现上述图像处理方法。
进一步地，所述计算机设备还包括电源和输入单元，所述电源通过电源管理系统与所述处理器逻辑相连。
有益效果
本发明图像处理方法、图像处理装置及计算机设备，根据待处理图片中是否含有预设场景，由对应的预设场景的相关性系数进行修正，调整根据高斯模型所获得的预设场景预设色彩的高斯概率，再由相应的补偿处理技术，例如采用视角补偿处理后，使得图片显示时，预设色彩的显示过渡自然，符合人眼视觉特性，能够有效改善色偏现象，还能够提高对待处理图片的处理效率，提高色彩侦测的准确性，以避免对其他场景中与预设场景预设色彩相近的色彩的误侦。同时，由于仅对待处理图片中的预设场景预设色彩进行处理，因此，在图片输出时，还能够有效降低格感，提高图片质量。
附图说明
为了更清楚地说明本发明实施例中的技术方案，下面将对实施例描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1是本发明实施例中提供的图像处理方法的示意图；
图2是本发明实施例中提供的图像处理方法的高斯模型模拟效果图；
图3是本发明实施例中提供的图像处理方法的对人像场景的肤色色彩处理的概率效果图；
图4是本发明实施例中提供的图像处理方法的对人像场景的肤色色彩、蓝天场景的蓝色色彩和草地场景的绿色色彩处理的概率效果图；
图5本发明实施例中一种计算机设备的结构示意图。
图中，1-电源，2-处理器，3-存储器，4-输入单元。
本发明的实施方式
下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。
在本发明的描述中，需要理解的是，术语“中心”、“纵向”、“横向”、“长度”、“宽度”、“厚度”、“上”、“下”、“前”、“后”、“左”、“右”、“竖直”、“水平”、“顶”、“底”、“内”、“外”等指示的方位或位置关系为基于附图所示的方位或位置关系，仅是为了便于描述本发明和简化描述，而不是指示或暗示所指的装置或元件必须具有特定的方位、以特定的方位构造和操作，因此不能理解为对本发明的限制。此外，术 语“第一”、“第二”仅用于描述目的，而不能理解为指示或暗示相对重要性或者隐含指明所指示的技术特征的数量。由此，限定有“第一”、“第二”的特征可以明示或者隐含地包括一个或者更多个特征。在本发明的描述中，“多个”的含义是两个或两个以上，除非另有明确具体的限定。
在本发明中，“示例性”一词用来表示“用作例子、例证或说明”。本发明中被描述为“示例性”的任何实施例不一定被解释为比其它实施例更优选或更具优势。为了使本领域任何技术人员能够实现和使用本发明，给出了以下描述。在以下描述中，为了解释的目的而列出了细节。应当明白的是，本领域普通技术人员可以认识到，在不使用这些特定细节的情况下也可以实现本发明。在其它实例中，不会对公知的结构和过程进行详细阐述，以避免不必要的细节使本发明的描述变得晦涩。因此，本发明并非旨在限于所示的实施例，而是与符合本发明所公开的原理和特征的最广范围相一致。如无特殊说明，本发明中所涉及的方位上的平行或垂直等，并不是严格意义上的平行或垂直，只要相应的结构能够实现相应的目的即可。
请参阅图1，本发明实施例提供一种图像处理方法，所述图像处理方法包括：
步骤S1、获取待处理图片中关于预设色彩的第一待处理色度数据集和第二待处理色度数据集；
步骤S2、根据第一待处理色度数据集和第二待处理色度数据集，获取预设场景的预设色彩的高斯概率；
步骤S3、识别待处理图片中含有预设场景的预设色彩的图片区域；
根据高斯概率，获取图片区域中预设场景的预设色彩的色彩输出修正值，以对待处理图片的预设场景中的预设色彩进行补偿处理。
具体地，本发明实施例中，由所构建的高斯模型对待处理图片中的相关数据进行处理，以获取待处理图片中关于预设场景预设色彩的高斯概率。在构建高斯模型时，可在相关的数据库选取多张预处理图片，这些预处理图片中分别包含相应的预设场景。预设场景的类型和 数量可根据实际需求具体设定，本发明实施例中的图片处理方法，在构建高斯模型时，预设场景可以人像、蓝天、草地、食物、动物、其他自然景色、建筑物等多种场景，相应的预设色彩是各场景中相应的色彩。例如，人像场景时，预设色彩可为肤色；蓝天场景时，预设色彩可为蓝色；草地场景时，预设色彩可为绿色。
本发明实施例中在构建高斯模型时，选取人眼敏感色以及对应的场景为例进行说明。本发明实施例中，选取人像、蓝天和草地三种预设场景，相应的预设色彩分别为肤色、蓝色和绿色。以下皆以上述三种预设场景以及相应的预设色彩为例进行说明。
在数据库中分别选取多张含有人像预设场景的第一预处理图片、多张含有蓝天预设场景的第二预处理图片和多张含有草地预设场景的第三预处理图片。含有各预设场景的预处理图片的数量根据实际情况具体设定。
对于任意一张第一预处理图片，提取该第一预处理图片的肤色数据，具体提取的方法采用目前常规的提取方法即可。在得到第一预处理图片的肤色数据后，可在Ycbcr空间将该肤色数据进行分解处理，分别得到关于该肤色数据的亮度数据以及第一初始色度数据和第二初始色度数据。
则对于多张第一预处图片的肤色数据，能够得到关于肤色数据的亮度数据集、第一初始色度数据集和第二初始色度数据集。其中，对肤色数据的分解处理，可以是在Ycbcr空间内处理，也可以是在HSB色彩空间等进行处理；同样地，以下对其他预设场景的预设色彩进行分解处理的方式可在Ycbcr空间内处理，也可才其他色彩空间进行处理，以下均以在Ycbcr空间内处理为例进行说明。
具体地，第一预处理图片的肤色数据可具体采用如下公式进行处理：
y skin(i)＝(R*0.2567+G*0.5041+B*0.0979)+16  (1)
cb skin(i)＝(R*0.1482+G*0.2909+B*0.4391)+128  (2)
cr skin(i)＝(R*0.4392+G*0.3678+B*0.0714)+128  (3)
其中，上述公式中的R、G、B分别为肤色数据的红色分量值、绿色分量值和蓝色分量值，y skin(i)为肤色数据的亮度数据，cb skin(i)为肤色数据的第一初始色度数据，cr skin(i)为肤色数据的第二初始色度数据。
分别对多张第一预处理图片进行上述同样的处理，得到多个亮度数据y skin，以形成亮度数据集；得到多个第一初始色度数据cb skin，得到第一初始色度数据集cb skin(1)、cb skin(2)......cb skin(i).......；得到多个第二初始色度数据cr skin，以形成第二初始色度数据集cr skin(1)、cr skin (2).......cr skin(i).......。
求取第一初始色度数据集的均值，得到关于肤色数据的第一色度均值μ skin1；并求取第一初始色度数据集中的各第一初始色度数据与上述第一色度均值μ skin1之间的方差a skin。求取第二初始色度数据集的均值，得到关于肤色数据的第二色度均值μ skin2；并求取第二初始色度数据集中的各第二初始色度数据与上述第二色度均值μ skin2之间的方差d skin。
由方差a skin、d skin、第一初始色度数据集cb skin(1)、cb skin(2)......cb skin (i).......，第二初始色度数据集cr skin(1)、cr skin(2).......cr skin(i).......，获取关于人像场景下肤色色彩的第一初始色度数据和第二初始色度数据的协方差矩阵cov(cb skin,cr skin)，具体表述如下：
Figure PCTCN2021100153-appb-000001
其中，cb skin(i)为任一第一预处理图片的第一初始色度数据，cr skin (i)为任一第一预处理图片的第二初始色度数据，μ skin1为多张第一预处理图片的肤色数据的第一色度均值，μ skin2为多张第一预处理图片的肤色数据的第二色度均值，a skin为第一预处理图片的肤色数据的第一初始色度数据与上述第一色度均值μ skin1之间的方差矩阵，d skin为第一预处理图片的肤色数据的第二初始色度数据与上述第二色度均值μ skin2之间的方差矩阵，b skin、c skin为关于第一预设图片的肤色色彩的第一初始色度数据集与第二初始色度数据集之间的相关度。
由上述公式(4)可获取关于cov(cb skin,cr skin)的逆矩阵cov -1 (cb skin,cr skin)或Σ skin -1，以及cov(cb skin,cr skin)的秩|Σ skin|。由上述各参数即可构建关于人像场景下肤色色彩的高斯模型，其具体可表述如下：
Figure PCTCN2021100153-appb-000002
其中，A为高斯模型的幅值，取值范围为[0，1]，gauss skin(cb i,cr i)为人像场景下高斯模型所得的关于肤色色彩的初始概率，a skin为第一预处理图片的肤色数据的第一初始色度数据与上述第一色度均值μ skin1之间的方差矩阵，d skin为第一预处理图片的肤色数据的第二初始色度数据与上述第二色度均值μ skin2之间的方差矩阵，cb i为关于肤色色彩的第一色度变量，cr i为关于肤色色彩的第二色度变量，Σ skin -1为cov(cb skin,cr skin)的逆矩阵，|Σ skin|为cov(cb skin,cr skin)的秩，μ skin为关于第一预设图片的肤色色彩的第一初始色度数据集与第二初始色度数据集的均值。
同样地，对于任意一张第二预处理图片，提取该第二预处理图片的蓝色数据。在得到第二预处理图片的蓝色数据后，可在Ycbcr空间将该蓝色数据进行分解处理，分别得到关于该蓝色数据的亮度数据、第一初始色度数据和第二初始色度数据。则对于多张第二预处理图片的蓝色数据，能够得到关于蓝色数据的亮度数据集、第一初始色度数据集和第二初始色度数据集。
具体地，第二预处理图片的蓝色数据可具体采用如下公式进行处理：
y sky(i)＝(R*0.2567+G*0.5041+B*0.0979)+16  (6)
cb sky(i)＝(R*0.1482+G*0.2909+B*0.4391)+128  (7)
cr sky(i)＝(R*0.4392+G*0.3678+B*0.0714)+128  (8)
其中，上述公式中的R、G、B分别为蓝色数据的红色分量值、绿色分量值和蓝色分量值，y sky(i)为蓝色数据的亮度数据，cb sky(i)为蓝色数据的第一初始色度数据，cr sky(i)为蓝色数据的第二初始色度数据。
由第二预处理图片的上述数据可获取关于蓝天场景下蓝色色彩 的第一初始色度数据和第二初始色度数据的协方差矩阵cov(cb sky,cr sky)，具体表述如下：
Figure PCTCN2021100153-appb-000003
Figure PCTCN2021100153-appb-000004
上述公式(9)和(10)中，cb sky(i)为任一第二预处理图片的第一初始色度数据，cr sky(i)为任一第二预处理图片的第二初始色度数据，μ sky1为多张第二预处理图片的蓝色数据的第一色度均值，μ sky2为多张第二预处理图片的蓝色数据的第二色度均值，asky为第二预处理图片的蓝色数据的第一初始色度数据与上述第一色度均值μsky1之间的方差矩阵，dsky为第二预处理图片的蓝色数据的第二初始色度数据与上述第二色度均值μsky2之间的方差矩阵，bsky、csky为第一初始色度数据集与第二初始色度数据集之间的相关度；A为高斯模型的幅值，取值范围为[0，1]，gausssky(cbi,cri)为蓝天场景下高斯模型所得的关于蓝色色彩的初始概率，asky为第二预处理图片的蓝色数据的第一初始色度数据与上述第一色度均值μsky1之间的方差矩阵，dsky为第二预处理图片的蓝色数据的第二初始色度数据与上述第二色度均值μsky2之间的方差矩阵，cbi为关于蓝色色彩的第一色度变量，cri为关于蓝色色彩的第二色度变量，Σsky-1为cov(cbsky,crsky)的逆矩阵，|Σsky|为cov(cbsky,crsky)的秩，
Figure PCTCN2021100153-appb-000005
为关于第二预设图片的蓝色色彩的第一初始色度数据集与第二初始色度数据集的均值。
对于任意一张第三预处理图片，提取该第三预处理图片的绿色数据。在得到第三预处理图片的绿色数据后，可在Ycbcr空间将该绿色数据进行分解处理，分别得到关于该绿色数据的亮度数据、第一初始色度数据和第二初始色度数据。则对于多张第三预处理图片的绿色数据，能够得到关于绿色数据的亮度数据集、第一初始色度数据集和第二初始色度数据集。
具体地，第三预处理图片的绿色数据可具体采用如下公式进行处理：
y grass(i)＝(R*0.2567+G*0.5041+B*0.0979)+16  (11)
cb grass(i)＝(R*0.1482+G*0.2909+B*0.4391)+128  (12)
cr grass(i)＝(R*0.4392+G*0.3678+B*0.0714)+128  (13)
其中，上述公式中的R、G、B分别为绿色数据的红色分量值、绿色分量值和蓝色分量值，y grass(i)为绿色数据的亮度数据，cb grass(i)为绿色数据的第一初始色度数据，cr grass(i)为绿色数据的第二初始色度数据。
由第二预处理图片的上述数据可获取关于蓝天场景下蓝色色彩的第一初始色度数据和第二初始色度数据的协方差矩阵cov(cb sky,cr sky)，具体表述如下：
Figure PCTCN2021100153-appb-000006
Figure PCTCN2021100153-appb-000007
上述公式(14)和(15)中，cb grass(i)为任一第三预处理图片的第一初始色度数据，cr grass(i)为任一第三预处理图片的第二初始色度数据，μ grass1为多张第三预处理图片的绿色数据的第一色度均值，μ grass2为多张第三预处理图片的绿色数据的第二色度均值，a grass为第三预处理图片的绿色数据的第一初始色度数据与上述第一色度均值μ grass1之间的方差矩阵，d grass为第三预处理图片的绿色数据的第二初始色度数据与上述第二色度均值μ grass2之间的方差矩阵，b grass、c grass为第一初始色度数据集与第二初始色度数据集之间的相关度；A为高斯模型的幅值，取值范围为[0，1]，gauss grass(cb i,cr i)为草地场景下高斯模型所得的关于绿色色彩的初始概率，a grass为第三预处理图片的绿色数据的第一初始色度数据与上述第一色度均值μ grass1之间的方差矩阵，d grass为第三预处理图片的绿色数据的第二初始色度数据与上述第二色度 均值μ grass2之间的方差矩阵，cb i为关于绿色色彩的第一色度变量，cr i为关于绿色色彩的第二色度变量，Σ grass-1为cov(cb grass,cr grass)的逆矩阵，|Σ grass|为cov(cb grass,cr grass)的秩，
Figure PCTCN2021100153-appb-000008
为关于第三预设图片的绿色色彩的第一初始色度数据集与第二初始色度数据集的均值。
采用上述方法构建高斯模型，能够根据需求设定预设场景的类型及数量，相应的设定预设色彩的类型及数量，分别建立各预设场景下各预设色彩的高斯模型，以根据不同应用场景、客户不同需求或图片质量要求等，灵活调整高斯模型的具体构成。并且，可根据需求调整高斯模型中的幅值、预处理图片中预设色彩的均值、相关协方差矩阵等参数还可以根据精度或其他考量而进行调整，其实用性和通用性强。
具体地，在对待处理图片进行处理时，先提取待处理图片的预设色彩的数据。例如，当对待处理图片中的人像、蓝天和草地场景进行处理时，分别提取待处理图片中的肤色、蓝色和草地的色彩数据，并分别在Ycbcr空间进行分解处理，以得到关于肤色色彩的第一待处理色度数据集和第二待处理色度数据集、关于蓝色色彩的第一待处理色度数据集和第二待处理色度数据集，以及关于绿色色彩的第一待处理色度数据集和第二待处理色度数据集。
获取各预设色彩的第一待处理色度数据集和第二待处理色度数据集后，即可将各预设色彩的第一待处理色度数据集以及第二待处理色度数据集中的各色度数据代入到相应的预设色彩的高斯模型中，以得到关于该预设色彩的初始概率图。
例如，将肤色色彩的第一待处理色度数据集和第二待处理色度数据集中的各色度数据对应代入到上述公式(5)中，即可获得待处理图片中关于肤色色彩的初始概率图。同样地，将蓝色色彩的第一待处理色度数据集和第二待处理色度数据集中的各色度数据对应代入到上述公式(10)中，即可获得待处理图片中关于蓝色色彩的初始概率图。将绿色色彩的第一待处理色度数据集和第二待处理色度数据集中的各色度数据对应代入到上述公式(15)中，即可获得待处理图片中关于绿色色彩的初始概率图。
具体地，在对待处理图片进行处理时，可对待处理图片中是否含有预设场景进行判断，并根据判断结果对预设场景的相关性系数进行赋值。根据待处理图片中是否含有预设场景，由对应的预设场景的相关性系数进行修正，调整根据高斯模型所获得的预设场景预设色彩的高斯概率，不仅能够提高对待处理图片的处理效率，还能够提高色彩侦测的准确性，以避免对其他场景中与预设场景预设色彩相近的色彩的误侦。同时，由于仅对待处理图片中的预设场景预设色彩进行处理，因此，在图片输出时，还能够有效降低格感，提高图片质量。其中，对待处理图片中是否含有预设场景进行判断的方法可采用目前常规的方式处理即可。
具体地，当设置有多个预设场景以及对应的预设色彩时，关于多个预设场景预设色彩的综合高斯概率值，其可由经各预设场景的相关性系数修正后的初始概率之和得到，具体可由如下公式获取：
gauss(cb,cr)＝α*gauss skin(cb i,cr i)+β*gauss sky(cb i,cr i)+γ*gauss grass(cb i,cr i)  (16)
式中，gauss(cb，cr)为关于待处理图片中预设场景预设色彩的高斯概率，α为待处理图片中人像场景的相关性系数，gauss skin(cb i,cr i)为高斯模型所得的关于肤色色彩的初始概率，β为待处理图片中蓝天场景的相关性系数，gauss sky(cb i,cr i)为由高斯模型所得的关于蓝色色彩的初始概率，γ为待处理图片中草地场景的相关性系数，gauss grass(cb i,cr i)为由高斯模型所得的关于绿色色彩的初始概率。
当待处理图片中没有相应的预设场景时，可对应地将相应的相关性系数赋值为0，则其与由高斯模型拟合得到的关于该预设场景预设色彩的初始概率之积为0，以避免对待处理图片中的类似色彩的误侦。
例如，当待处理图片中含有人像场景时，对关于人像场景的相关性系数α赋值为1；当待处理图片中没有人像场景时，对关于人像场景的相关性系数α赋值为0。同样地，当待处理图片中含有蓝天场景时，对关于蓝天场景的相关性系数β赋值为1；当待处理图片中没有蓝天场景时，对关于蓝天场景的相关性系数β赋值为0。当待处理图片中含有 草地场景时，对关于草地场景的相关性系数γ赋值为1；当待处理图片中没有草地场景时，对关于草地场景的相关性系数γ赋值为0。
例如，当待处理图片中含有人像场景，没有蓝天场景和草地场景时，待处理图片的色彩数据经高斯模型拟合，并经相应的预设场景相关性系数修正后，所得的高斯概率为gauss(cb,cr)＝gauss skin(cb i,cr i)，对于蓝天场景和草地场景的高斯拟合概率为0。当待处理图片中含有人像场景、蓝天场景，没有草地场景时，待处理图片的色彩数据经高斯模型拟合后，并经相应的预设场景相关性系数修正后，所得的高斯概率为gauss(cb,cr)＝gauss skin(cb i,cr i)+gauss sky(cb i,cr i)。当待处理图片中同时含有人像场景、蓝天场景和草地场景时，待处理图片的色彩数据经高斯模型拟合后，并经相应的预设场景相关性系数修正后，所得的高斯概率为gauss(cb,cr)＝gauss skin(cb i,cr i)+gauss sky(cb i,cr i)+gauss grass(cb i,cr i)，参见图2中(a)和(b)图，分别为高斯拟合模型数据模拟图的正视效果图和俯视效果图。
通常情况下，待处理图片中除了含有人像场景、蓝天场景和/或草地场景时，通常还会存在其他场景，例如图3中(a)图中人像背后的物品柜等。在对待处理图片进行处理时，还可通过识别待处理图片中含有预设场景的预设色彩的图片区域，以更准确地对待处理图片中预设场景预设颜色进行模拟处理。识别到该部分图片区域后，应用由高斯模型所获取的高斯概率，即能够获取到待处理图片中关于预设场景预设色彩的色彩输出修正值。该色彩输出修正值应用到视角补偿技术中，能够有效改善输出的图片中关于预设场景预设色彩的色偏现象。并且，由高斯模型拟合处理后的色彩数据，其色彩过渡自然，格感小，图片质量更高，符合人眼视觉特性，观感体验更佳。
具体地，在识别待处理图片中预设场景的图片区域时，可先将待处理图片划分为多个图片区域，图片区域的大小可根据实际情况或需求具体调整，例如，可将待处理图片划分为10*10大小的区域。
对于任一图片区域，提取该图片区域内的像素值，获取该图片区域中的像素梯度；例如，可采用sobel算子等目前常规算法以计算该图 片区域的像素梯度T。设定第一预设阈值T thresh，对各像素处的像素梯度值与该第一预设阈值T thresh进行比较。对于高于第一预设阈值T thresh的像素梯度值，对与该像素梯度值对应的像素进行计数。
统计该图片区域内像素梯度高于第一预设阈值T thresh的像素数量l，该图片区域内像素梯度值高于第一预设阈值T thresh的像素数量l与第二预设阈值L thresh进行比较。若该图片区域内像素梯度值高于第一预设阈值T thresh的像素数量l大于第二预设阈值L thresh，则判定该图片区域为细纹理区，该图片区域内含有预设场景的预设色彩。其中，第一预设阈值T thresh和第二预设阈值L thresh均可根据实际情况或需求具体设定。
具体地，对预设场景预设色彩进行更准确识别时，对含有预设场景预设色彩的图片区域和没有含有预设场景预设色彩的图片区域分别赋值，得到含有预设场景预设色彩的图片区域的第一纹理概率P 1，并赋值为1；得到没有含有预设场景预设色彩的图片区域的第二纹理概率P 2，并赋值为0。按照上述方式，对待处理图片的各图片区域进行识别，并根据识别的情况分别赋值。
在识别待处理图片中含有预设场景的预设色彩的图片区域后，作为其中一种获取色彩输出修正值的实现方式，对于任一图片区域，该图片区域相应的高斯概率与纹理概率之积，获得该图片区域关于预设场景预设色彩的色彩输出修正值。则当该图片区域含有预设场景预设色彩时，该图片区域对应色色彩输出修正值不为0，该色彩输出修正值用于视角补偿技术，以改善色偏现象；当该图片区域不含有预设场景预设色彩时，该图片区域对应的色彩输出修正值为0，则对该图片区域的色彩不作处理。
由于人像场景、蓝天场景和草地场景均为细纹理区，通过上述处理方法，能够在对人像场景等预设场景进行初步判断的基础上，能够进一步识别人像场景等预设场景所在的具体区域，避免对与预设场景中预设色彩相似的色彩等情形出现误侦，有效提升对预设场景预设颜色等数据提取的准确性，降低过检率和误检率。
在识别待处理图片中含有预设场景的预设色彩的图片区域后，作 为另一种获取色彩输出修正值的实现方式，则该图片区域对应的纹理概率P，赋值为1。并且，仅针对含有预设场景的预设色彩的图片区域，提取相应的色彩数据，并采用高斯模型拟合处理，得到该图片区域对应的高斯概率，该高斯概率即为该图片区域对应的色彩输出修正值。
具体地，色彩输出修正值可具体用如下算法实现：
gauss out＝gauss(cb,cr)*P  (17)
上式中，gauss out为关于待处理图片中预设场景预设色彩的色彩输出修正值，gauss(cb，cr)为关于待处理图片中预设场景预设色彩的高斯概率，P为待处理图片中各图片区域关于预设场景预设色彩的纹理概率，或者，为待处理图片中含有预设场景预设色彩的图片区域的纹理概率。
采用该方式进行处理时，无需对没有含有预设场景预设色彩的图片区域采集数据，也无需针对该一部分区域进行高斯模型拟合处理，能够简化数据处理过程，降低数据处理复杂程度。
进一步地，在对待处理图片进行处理时，还可根据待处理图片的预设场景的预设色彩的亮度数据进一步提高对待处理图片预设场景预设色彩侦测的准确性。
具体地，可根据亮度数据将预设场景的预设色彩的亮度划分为多个亮度区间，不同的亮度区间内，采用不同的线性调节模型分别对对应的预设色彩的色彩数据的亮度数据进行修正或抑制调节，能够降低对于低灰阶图片数据采集中可能出现的闪烁问题，能够进一步提高对预设场景预设色彩的色彩数据处理的准确性，参见图3的(a)是原图，图3的(b)图对图(a)采用本发明实施例的图像处理方法对人像场景的肤色色彩侦测处理的概率效果图，图3的(c)是(b)图的细节图，其中，图3的(a)、(b)和(c)图中的d框、e框和f框所指代的位置分别一一对应。图4的(a)图是原图，图4的(b)图是对(a)图采用本发明实施例的图像处理方法对人像场景的肤色色彩、蓝天场景的蓝色色彩和草地场景的绿色色彩侦测处理的概率效果图。
其中，线性调节模型可具体表述如下：
Figure PCTCN2021100153-appb-000009
上述公式(18)中，K(y)为关于预设场景预设色彩的亮度数据经线性调节模型修正后的亮度调整概率，y为预设场景预设色彩的亮度数据，k 1、k 2、k 3、k 4、l 1、l 2、l 3、l 4分别为设定参数。
其中，关于k 1、k 2、k 3、k 4、l 1、l 2、l 3、l 4，可分别根据设置不同亮度区间的概率设定；在实际应用中，还可根据所应用的面板的实际亮度等情况进行调整。例如，k 1、k 2、k 3、k 4、l 1、l 2、l 3、l 4可分别按照如下赋值进行初始设定，则线性调节模型可表述为：
Figure PCTCN2021100153-appb-000010
待处理图片经线性调节模型进行亮度抑制调节后，所获得的亮度调整概率与待处理图片中的预设场景预设色彩的高斯概率，以及待处理图片中关于预设场景预设色彩的图片区域的纹理概率P，综合对待处理图片进行处理，能够有效提高数据处理的准确性，降低过检和误检率，其具体可采用如下算法实现：
gauss out＝gauss(cb,cr)*P*K(y)  (20)
其中，gauss out为关于待处理图片中预设场景预设色彩的色彩输出修正值，gauss(cb，cr)为关于待处理图片中预设场景预设色彩的高斯概率，P为待处理图片中各图片区域关于预设场景预设色彩的纹理概率，或者，为待处理图片中含有预设场景预设色彩的图片区域的纹理概率，K(y)为关于预设场景预设色彩的亮度数据经线性调节模型修正后的亮度调整概率。
获取待处理图片中预设场景预设色彩的色彩输出修正值后，可将该色彩输出修正值应用于视角补偿技术中，具体实现过程，本发明实施例不做具体限定，例如，其处理过程可简述如下：
将待处理图片的RGB像素值展开，由m*n的待处理图片长宽比变成m*(n*3)的图片I RGB长宽比；
将展开图按s＝6个子像素作为一个分区，一共有ss＝(n*3)/s个分区；
若m为奇数行，且mod(s/3)＝1时，索引R像素的table值，V table＝H R；
若m为偶数行，且mod(s/3)＝1时，索引R像素的table值，V table＝L R；
若m为奇数行，且mod(s/3)＝2时，索引G像素的table值，V table＝H G；
若m为偶数行，且mod(s/3)＝2时，索引G像素的table值，V table＝L G；
若m为奇数行，且mod(s/3)＝0时，索引B像素的table值，V table＝H B；
若m为偶数行，且mod(s/3)＝0时，索引B像素的table值，V table＝L R；
计算视角补偿(View Angle Compensation，VAC)算法的输出结果：
out＝I RGB+gauss out*(V table-I RGB)  (21)
其中，V table为通过索引视角补偿技术中驱动模式的内置表格而设定的驱动电压，I RGB为待处理图片的RGB值，L R、H R、L G、L B、H B分别为视角补偿技术中驱动模式的内置表格中RGB通道的设定值。
本发明实施例还提供一种图像处理装置，包括：
数据采集单元，用于获取待处理图片中关于预设色彩的第一待处理色度数据集和第二待处理色度数据集；
第一数据处理单元，用于根据第一待处理色度数据集和第二待处理色度数据集，由所构建的高斯模型获取预设场景的预设色彩的高斯概率；
纹理识别单元，用于识别待处理图片中含有预设场景的图片区域；
第二数据处理单元，用于根据高斯概率，获取图片区域中预设场景的预设色彩的色彩输出修正值，以对待处理图片的预设场景中的预 设色彩进行补偿处理。
其中，根据预处理图片的预设场景中预设色彩的第一初始色度数据集和第二初始色度数据集，建立关于预设色彩的高斯模型。
具体地，该图像处理装置还包括亮度调节单元，用于获取待处理图片中关于预设色彩的亮度数据；判断亮度数据所在的亮度区间；调取亮度区间的线性调节模型；根据线性调节模型和亮度数据，获取待处理图片中关于亮度的亮度调整概率。此时，数据处理单元，用于根据亮度调整概率和高斯概率，获取图片区域中预设场景的预设色彩的色彩输出修正值，以对待处理图片的预设场景中的预设色彩进行补偿处理。
该图像处理装置，由所构建的高斯模型对预设场景预设色彩进行拟合处理，并根据待处理图片中是否含有预设场景，由对应的预设场景的相关性系数进行修正，调整根据高斯模型所获得的预设场景预设色彩的高斯概率，在对图片进行视角补偿处理后，使得图片显示时，预设色彩的显示过渡自然，符合人眼视觉特性，能够有效改善色偏现象，还能够提高对待处理图片的处理效率，提高色彩侦测准确性，以避免对其他场景中与预设场景预设色彩相近的色彩的误侦。并且，由高斯模型拟合处理后的色彩数据，其色彩过渡自然，格感小，图片质量更高，符合人眼视觉特性，观感体验更佳。
本发明实施例还提供一种计算机设备，包括：
一个或多个处理器；
存储器；以及
一个或多个应用程序，其中一个或多个应用程序被存储于存储器中，并配置为由处理器执行以实现上述图像处理方法。
该计算机设备可以是独立的服务器，也可以是服务器组成的服务器网络或服务器集群，例如，本发明实施例中所描述的计算机设备，其包括但不限于计算机、网络主机、单个网络服务器、多个网络服务器集或多个服务器构成的云服务器。其中，云服务器由基于云计算(Cloud Computing)的大量计算机或网络服务器构成。
可以理解的是，本发明实施例中所使用的计算机设备可以是既包括接收和发射硬件的设备，即具有能够在双向通信链路上，执行双向通信的接收和发射硬件的设备。这种设备可以包括：蜂窝或其他通信设备，其具有单线路显示器或多线路显示器或没有多线路显示器的蜂窝或其他通信设备。具体的计算机设备具体可以是台式终端或移动终端，计算机设备具体还可以是手机、平板电脑、笔记本电脑等中的一种。
该计算机设备可以包括一个或者一个以上处理核心的处理器2、一个或一个以上的存储器3、电源1和输入单元4等部件。本领域技术人员可以理解，图5中示出的计算机设备结构并不构成对计算机设备的限定，可以包括比图示更多或更少的部件，或者组合某些部件，或者不同的部件布置。其中：
处理器2是该计算机设备的控制中心，利用各种接口和线路连接整个计算机设备的各个部分，通过运行或执行存储在存储器3内的目标文件和/或模块，以及调用存储在存储器3内的目标文件，执行计算机设备的各种功能和处理数据，从而对计算机设备进行整体监控。可选的，处理器2可包括一个或多个处理核心；优选的，处理器2可集成应用处理器和调制解调处理器，其中，应用处理器主要处理操作系统、用户界面和应用程序等，调制解调处理器主要处理无线通信。可以理解的是，上述调制解调处理器也可以不集成到处理器2中。
存储器3可用于存储软件程序等目标文件以及模块，处理器2通过运行存储在存储器3的目标文件以及模块，从而执行各种功能应用以及数据处理。存储器3可主要包括存储程序区和存储数据区，其中，存储程序区可存储操作系统、至少一个功能所需的应用程序(比如声音播放功能、图像播放功能等)等；存储数据区可存储根据计算机设备的使用所创建的数据等。此外，存储器3可以包括高速随机存取存储器，还可以包括非易失性存储器，例如至少一个磁盘存储器件、闪存器件、或其他易失性固态存储器件。相应地，存储器3还可以包括存储器控制器，以提供处理器2对存储器3的访问。
计算机设备还包括给各个部件供电的电源1，优选的，电源1可以 通过电源管理系统与处理器2逻辑相连，从而通过电源管理系统实现管理充电、放电、以及功耗管理等功能。电源1还可以包括一个或一个以上的直流或交流电源、再充电系统、电源故障检测电路、电源转换器或者逆变器、电源状态指示器等任意组件。
该计算机设备还可包括输入单元4，该输入单元4可用于接收输入的数字或字符信息，以及产生与用户设置以及功能控制有关的键盘、鼠标、操作杆、光学或者轨迹球信号输入。
尽管未示出，计算机设备还可以包括显示单元等，在此不再赘述。具体在本实施例中，计算机设备中的处理器2会按照如下的指令，将一个或一个以上的应用程序的进程对应的可执行文件加载到存储器3中，并由处理器2来运行存储在存储器3中的目标文件。
本发明的图像处理方法、图像处理装置及计算机设备，该图片处理方法由所构建的高斯模型对预设场景预设色彩进行拟合处理，并根据待处理图片中是否含有预设场景，由对应的预设场景的相关性系数进行修正，调整根据高斯模型所获得的预设场景预设色彩的高斯概率，在对图片进行视角补偿处理后，使得图片显示时，预设色彩的显示过渡自然，符合人眼视觉特性，能够有效改善色偏现象，还能够提高对待处理图片的处理效率，提高色彩侦测的准确性，以避免对其他场景中与预设场景预设色彩相近的色彩的误侦。同时，由于仅对待处理图片中的预设场景预设色彩进行处理，因此，在图片输出时，还能够有效降低格感，提高图片质量。
在上述实施例中，对各个实施例的描述都各有侧重，某个实施例中没有详述的部分，可以参见上文针对其他实施例的详细描述，此处不再赘述。
具体实施时，以上各个单元或结构可以作为独立的实体来实现，也可以进行任意组合，作为同一或若干个实体来实现，以上各个单元或结构的具体实施可参见前面的方法实施例，在此不再赘述。
以上各个操作的具体实施可参见前面的实施例，在此不再赘述。
本发明中应用了具体个例对本发明的原理及实施方式进行了阐 述，以上实施例的说明只是用于帮助理解本发明的方法及其核心思想；同时，对于本领域的技术人员，依据本发明的思想，在具体实施方式及应用范围上均会有改变之处，综上所述，本说明书内容不应理解为对本发明的限制。

Claims (20)

  1. 一种图像处理方法，其中，所述图像处理方法包括：
    获取待处理图片中关于预设色彩的第一待处理色度数据集和第二待处理色度数据集；
    根据所述第一待处理色度数据集和所述第二待处理色度数据集，获取所述预设场景的所述预设色彩的高斯概率；
    识别所述待处理图片中含有所述预设场景的所述预设色彩的图片区域；
    根据所述高斯概率，获取所述图片区域中所述预设场景的所述预设色彩的色彩输出修正值，以对所述待处理图片的所述预设场景中的所述预设色彩进行补偿处理。
  2. 根据权利要求1所述的图像处理方法，其中，所述图像处理方法还包括：
    获取所述待处理图片中关于所述预设色彩的亮度数据；
    判断所述亮度数据所在的亮度区间；
    调取所述亮度区间的线性调节模型；
    根据所述线性调节模型和所述亮度数据，获取所述待处理图片中关于亮度的亮度调整概率；
    根据所述亮度调整概率和所述高斯概率，获取所述图片区域中所述预设场景的所述预设色彩的色彩输出修正值，以对所述待处理图片的所述预设场景中的所述预设色彩进行补偿处理。
  3. 根据权利要求1所述的图像处理方法，其中，在所述根据所述第一待处理色度数据集和所述第二待处理色度数据集，获取所述预设场景的预设色彩的高斯概率之前，所述方法还包括：
    获取预处理图片的预设场景中预设色彩的第一初始色度数据集和第二初始色度数据集；
    根据所述第一初始色度数据集和所述第二初始色度数据集，建立关于所述预设色彩的高斯模型；
    所述根据所述第一待处理色度数据集和所述第二待处理色度数据集，获取所述预设场景的预设色彩的高斯概率，包括：
    根据所述第一待处理色度数据集和所述第二待处理色度数据集，由所述高斯模型获取所述预设场景的所述预设色彩的高斯概率。
  4. 根据权利要求3所述的图像处理方法，其中，获取预处理图片的预设场景中预设色彩的第一初始色度数据集和第二初始色度数据集，包括：
    获取多个含有所述预设场景的预处理图片；
    提取任一所述预处理图片中关于所述预设场景的所述预设色彩的色彩数据，获取所述第一初始色度数据集和所述第二初始色度数据集。
  5. 根据权利要求3所述的图像处理方法，其中，根据所述第一初始色度数据集和所述第二初始色度数据集，建立关于所述预设色彩的高斯模型，包括：
    分别获取所述第一初始色度数据集和所述第二初始色度数据集的均值；
    获取关于所述第一初始色度数据集和所述第二初始色度数据集的协方差矩阵、协方差矩阵的逆以及协方差矩阵的秩；
    根据所述协方差矩阵、所述协方差矩阵的逆以及所述协方差矩阵的秩，建立所述高斯模型。
  6. 根据权利要求1所述的图像处理方法，其中，所述预设场景包括人像、蓝天、草地、食物、动物和建筑物中的一种或多种。
  7. 根据权利要求3所述的图像处理方法，其中，所述获取预处理图片的预设场景中预设色彩的第一初始色度数据集和第二初始色度数据集，包括：
    获取多个含有人像预设场景的第一预处理图片；
    提取多个所述第一预处理图片的肤色数据；
    对所述肤色数据进行分解处理，得到所述肤色数据的第一初始色度数据集和第二初始色度数据集。
  8. 根据权利要求1所述的图像处理方法，其中，所述获取待处理图片中关于预设色彩的第一待处理色度数据集和第二待处理色度数据集，包括：
    提取待处理图片的预设色彩的数据；
    对所述预设色彩的数据进行分解处理；
    获取所述预设色彩的第一待处理色度数据集和第二待处理色度数据集。
  9. 根据权利要求1所述的图像处理方法，其中，获取所述预设场景的所述预设色彩的高斯概率，包括：
    判断所述待处理图片中是否含有所述预设场景；
    根据所述待处理图片中含有所述预设场景的判断结果，对所述待处理图片中关于所述预设场景的相关性系数赋值；
    对于所述待处理图片中的任一所述预设色彩，根据所述高斯模型获取所述预设色彩的初始概率；
    根据所述初始概率以及所述相关性系数，获取所述待处理图片中所述预设场景的高斯概率。
  10. 根据权利要求9所述的图像处理方法，其中，所述预设场景的数量有多个，所述预设色彩有多个；
    对于任一所述预设场景，所述预设场景的初始概率经相关性系数修正后，
    求取所述待处理图片中多个所述预设场景的经所述相关性系数修正后的所述初始概率之和，获取所述待处理图片中关于多个所述预设场景的高斯概率。
  11. 根据权利要求9所述的图像处理方法，其中，所述获取所述预设场景的所述预设色彩的高斯概率，还包括：
    判断所述待处理图片中是否含有所述预设场景；
    若所述待处理图片中没有所述预设场景，对所述待处理图片中关于所述预设场景的相关性系数赋值为0。
  12. 根据权利要求1所述的图像处理方法，其中，识别所述待处 理图片中含有所述预设场景的所述预设色彩的图片区域，包括：
    将所述待处理图片划分为多个所述图片区域；
    获取任一所述图片区域的像素梯度；
    统计任一所述图片区域中像素梯度高于第一预设阈值的像素数量；
    判断所述像素数量是否大于第二预设阈值；
    若所述像素数量大于第二预设阈值，所述图片区域为含有所述预设场景所述预设色彩的图片区域。
  13. 根据权利要求12所述的图像处理方法，其中，所述获取任一所述图片区域的像素梯度，包括：
    提取任一所述图片区域内的像素值；
    根据所述像素值计算所述图片区域的像素梯度。
  14. 根据权利要求12所述的图像处理方法，其中，根据所述高斯概率，获取所述图片区域中所述预设场景的所述预设色彩的色彩输出修正值，包括：
    若所述像素数量大于第二预设阈值，获取所述图片区域的第一纹理概率；
    若所述像素数量小于第二预设阈值，获取所述图片区域的第二纹理概率；
    对于任一所述图片区域，获取所述图片区域关于所述预设场景的所述预设色彩的高斯概率与所述第一纹理概率或第二纹理概率之积，得到所述图片区域中所述预设场景的所述预设色彩的色彩输出修正值。
  15. 根据权利要求12所述的图像处理方法，其中，根据所述高斯概率，获取所述图片区域中所述预设场景的所述预设色彩的色彩输出修正值，包括：
    获取含有所述预设场景的所述预设色彩的图片区域的第一纹理概率；
    获取所述第一纹理概率与含有所述预设场景的图片区域的所述 高斯概率之积，得到所述图片区域中所述预设场景的所述预设色彩的色彩输出修正值。
  16. 根据权利要求1所述的图像处理方法，其中，所述识别所述待处理图片中含有所述预设场景的所述预设色彩的图片区域，包括：
    获取所述待处理图片的亮度数据；
    根据所述亮度数据将所述预设场景的所述预设色彩的亮度划分为多个亮度区间；
    对所述亮度区间内的所述预设色彩的亮度数据进行修正或抑制调节，得到亮度调整概率。
  17. 一种图像处理装置，其中，包括：
    数据采集单元，用于获取待处理图片中关于预设色彩的第一待处理色度数据集和第二待处理色度数据集；
    第一数据处理单元，用于根据所述第一待处理色度数据集和所述第二待处理色度数据集，由所构建的高斯模型获取预设场景的所述预设色彩的高斯概率；
    纹理识别单元，用于识别所述待处理图片中含有所述预设场景的图片区域；
    第二数据处理单元，用于根据所述高斯概率，获取所述图片区域中所述预设场景的所述预设色彩的色彩输出修正值，以对所述待处理图片的所述预设场景中的所述预设色彩进行补偿处理。
  18. 根据权利要求17所述的图像处理装置，其中，所述图像处理装置还包括亮度调节单元，所述亮度调节单元用于获取所述待处理图片中关于所述预设色彩的亮度数据。
  19. 一种计算机设备，其中，包括：
    一个或多个处理器；
    存储器；以及
    一个或多个应用程序，其中所述一个或多个应用程序被存储于所述存储器中，并配置为由所述处理器执行以实现权利要求1-8中任一项所述的图像处理方法。
  20. 根据权利要求19所述的计算机设备，其中，所述计算机设备还包括电源和输入单元，所述电源通过电源管理系统与所述处理器逻辑相连。
PCT/CN2021/100153 2021-01-19 2021-06-15 图像处理方法、图像处理装置及计算机设备 WO2022156129A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US17/600,328 US20230105393A1 (en) 2021-01-19 2021-06-15 Image processing method, image processing device, and computer device

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110066060.3A CN112907457A (zh) 2021-01-19 2021-01-19 图像处理方法、图像处理装置及计算机设备
CN202110066060.3 2021-01-19

Publications (1)

Publication Number Publication Date
WO2022156129A1 true WO2022156129A1 (zh) 2022-07-28

Family

ID=76114975

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/100153 WO2022156129A1 (zh) 2021-01-19 2021-06-15 图像处理方法、图像处理装置及计算机设备

Country Status (3)

Country Link
US (1) US20230105393A1 (zh)
CN (1) CN112907457A (zh)
WO (1) WO2022156129A1 (zh)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112907457A (zh) * 2021-01-19 2021-06-04 Tcl华星光电技术有限公司 图像处理方法、图像处理装置及计算机设备
CN114913971B (zh) * 2022-06-10 2023-05-09 奇医天下大数据科技(珠海横琴)有限公司 一种基于人工智能的电子处方服务管理系统
US11961446B2 (en) 2022-06-15 2024-04-16 Shenzhen China Star Optoelectronics Semiconductor Display Technology Co., Ltd. Display driving method and device, and display device
CN115188308A (zh) * 2022-06-15 2022-10-14 深圳市华星光电半导体显示技术有限公司 显示装置的显示驱动方法及装置、显示装置

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102752477A (zh) * 2011-04-18 2012-10-24 三星电子株式会社 图像补偿设备、图像处理装置及其方法
CN103854261A (zh) * 2014-02-25 2014-06-11 上海理工大学 色偏图像的校正方法
EP3001668A1 (en) * 2014-09-24 2016-03-30 Thomson Licensing Method for compensating for color differences between different images of a same scene
CN107424179A (zh) * 2017-04-18 2017-12-01 微鲸科技有限公司 一种图像均衡方法及装置
US20180316927A1 (en) * 2017-04-26 2018-11-01 Google Inc. Chromatic aberration modeling in image compression and enhancement
CN109448653A (zh) * 2018-12-21 2019-03-08 惠科股份有限公司 显示装置补偿计算的方法、补偿结构及显示装置
CN112907457A (zh) * 2021-01-19 2021-06-04 Tcl华星光电技术有限公司 图像处理方法、图像处理装置及计算机设备

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2010014067A1 (en) * 2008-07-31 2010-02-04 Hewlett-Packard Development Company, L.P. Perceptual segmentation of images
JPWO2012005108A1 (ja) * 2010-07-09 2013-09-02 オリンパスメディカルシステムズ株式会社 画像記録再生システム
JP2014078808A (ja) * 2012-10-09 2014-05-01 Toshiba Corp 画像処理装置、画像処理方法
JP2015011585A (ja) * 2013-06-28 2015-01-19 株式会社リコー 画像処理装置、画像形成装置、画像形成システム、画像処理方法およびプログラム
CN104299196A (zh) * 2014-10-11 2015-01-21 京东方科技集团股份有限公司 一种图像处理装置及方法、显示设备
CN105678813A (zh) * 2015-11-26 2016-06-15 乐视致新电子科技（天津）有限公司 一种肤色检测方法及装置
CN109618173B (zh) * 2018-12-17 2021-09-28 深圳Tcl新技术有限公司 视频压缩方法、装置和计算机可读存储介质
WO2021016896A1 (zh) * 2019-07-30 2021-02-04 深圳市大疆创新科技有限公司 图像处理方法、系统、设备、可移动平台和存储介质

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102752477A (zh) * 2011-04-18 2012-10-24 三星电子株式会社 图像补偿设备、图像处理装置及其方法
CN103854261A (zh) * 2014-02-25 2014-06-11 上海理工大学 色偏图像的校正方法
EP3001668A1 (en) * 2014-09-24 2016-03-30 Thomson Licensing Method for compensating for color differences between different images of a same scene
CN107424179A (zh) * 2017-04-18 2017-12-01 微鲸科技有限公司 一种图像均衡方法及装置
US20180316927A1 (en) * 2017-04-26 2018-11-01 Google Inc. Chromatic aberration modeling in image compression and enhancement
CN109448653A (zh) * 2018-12-21 2019-03-08 惠科股份有限公司 显示装置补偿计算的方法、补偿结构及显示装置
CN112907457A (zh) * 2021-01-19 2021-06-04 Tcl华星光电技术有限公司 图像处理方法、图像处理装置及计算机设备

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
ZHANG TAO, WU HONGYING, LI MENG, LEI ZHI-CHUN: "A Novel Local Dimming Algorithm to Improve Image Contrast and Visual Quality", CHINESE JOURNAL OF ENGINEERING, vol. 39, no. 12, 31 December 2017 (2017-12-31), pages 1888 - 1897, XP055954266, ISSN: 2095-9389 *

Also Published As

Publication number Publication date
CN112907457A (zh) 2021-06-04
US20230105393A1 (en) 2023-04-06

Similar Documents

Publication Publication Date Title
WO2022156129A1 (zh) 图像处理方法、图像处理装置及计算机设备
EP3614333B1 (en) Image processing method, storage medium, and electronic apparatus
US7652717B2 (en) White balance correction in digital camera images
WO2019100675A1 (zh) 一种分区背光控制方法及其装置
US20200412942A1 (en) Photographing Method and Device, and Related Electronic Apparatus
CN109686342B (zh) 一种图像处理方法及装置
WO2019101005A1 (zh) 像素补偿方法、装置和终端设备
CN108337496A (zh) 白平衡处理方法、处理装置、处理设备及存储介质
DE102016122790A1 (de) Unterstützter automatischer Weißabgleich
US8830251B2 (en) Method and system for creating an image
US20180005409A1 (en) System and method facilitating palette-based color editing
CN105681775A (zh) 一种白平衡方法和装置
WO2017143852A1 (zh) 一种图片处理方法、装置及电子设备
CN111739041A (zh) 图像边框的裁剪方法、装置及设备
CN111062993A (zh) 色彩合并的作画图像处理方法、装置、设备及存储介质
CN113132696A (zh) 图像色调映射方法、装置、电子设备和存储介质
CN106600524A (zh) 一种图像处理方法及终端
WO2017152822A1 (zh) 图像处理方法和装置
CN107027017A (zh) 一种图像白平衡的调整方法、装置、图像处理芯片及存储装置
DE102018130083A1 (de) Schätzung der Beleuchtungsfarbigkeit beim automatischen Weißabgleich
CN113284063A (zh) 图像处理方法、图像处理装置、电子设备和可读存储介质
WO2010128683A1 (en) Blue sky color detection technique
US10026201B2 (en) Image classifying method and image displaying method
WO2023273111A1 (zh) 一种图像处理方法、装置、计算机设备和存储介质
CN114092359A (zh) 花屏处理方法、装置和电子设备

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21920511

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21920511

Country of ref document: EP

Kind code of ref document: A1