WO2019153799A1 - Procédé et appareil de compensation de distorsion des couleurs, et téléviseur - Google Patents

Procédé et appareil de compensation de distorsion des couleurs, et téléviseur Download PDF

Info

Publication number
WO2019153799A1
WO2019153799A1 PCT/CN2018/112335 CN2018112335W WO2019153799A1 WO 2019153799 A1 WO2019153799 A1 WO 2019153799A1 CN 2018112335 W CN2018112335 W CN 2018112335W WO 2019153799 A1 WO2019153799 A1 WO 2019153799A1
Authority
WO
WIPO (PCT)
Prior art keywords
ribbon
blocks
value
feature
feature blocks
Prior art date
Application number
PCT/CN2018/112335
Other languages
English (en)
Chinese (zh)
Inventor
黄哲
肖志林
Original Assignee
深圳创维－Rgb电子有限公司
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 深圳创维－Rgb电子有限公司 filed Critical 深圳创维－Rgb电子有限公司
Publication of WO2019153799A1 publication Critical patent/WO2019153799A1/fr

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N9/00—Details of colour television systems
    • H04N9/64—Circuits for processing colour signals
    • H04N9/646—Circuits for processing colour signals for image enhancement, e.g. vertical detail restoration, cross-colour elimination, contour correction, chrominance trapping filters

Definitions

  • the present disclosure relates to the field of digital video processing technologies, for example, to a method, apparatus, and television for compensating for color distortion.
  • the traditional de-striping technology mainly performs signal statistical processing on the small loss of the original color depth, and performs 2-4 times color depth compensation, which cannot be dealt with for serious color depth loss problems.
  • Embodiments of the present disclosure provide a method, an apparatus, and a television for compensating for color distortion, which can improve the reliability of compensating for color distortion.
  • Embodiments of the present disclosure provide a method for compensating for color distortion, including:
  • the image to be processed is input into the ribbon recognition neural network, and the plurality of ribbon feature blocks are obtained, including:
  • the convolution layer in the ribbon recognition neural network extracts a plurality of suspected ribbon blocks in the image to be processed; wherein the plurality of suspected ribbon regions The pixels in each of the suspected ribbon feature blocks are consecutive and the chrominance values of the pixels in each of the suspected ribbon feature blocks are the same;
  • the plurality of suspected ribbon blocks are analyzed using an activation function
  • the analyzing the plurality of suspected ribbon blocks by using an activation function comprises:
  • determining that the plurality of suspected ribbon blocks are a plurality of ribbon feature blocks including:
  • the number of ribbon feature blocks adjacent to each of the ribbon feature blocks is plural; and each of the plurality of ribbon feature blocks is characterized Obtaining a chrominance value of the block and a chrominance value of the adjacent color band feature block, obtaining a chromaticity compensation evaluation value of each of the color band characteristic blocks, comprising: separately calculating the plurality of color band characteristic blocks a difference between a chrominance value of each of the ribbon feature blocks and a chromaticity value of the adjacent ribbon feature block; calculating an average of the plurality of the difference values to obtain a color of the current ribbon feature block Degree compensation evaluation value.
  • determining a target chromaticity of each pixel in each of the ribbon feature blocks according to a chrominance compensation evaluation value of each of the plurality of ribbon feature blocks Values including:
  • a target chromaticity value for each pixel in each of the ribbon feature blocks is calculated based on a weight value of each pixel of each of the ribbon feature blocks.
  • determining a target color of each pixel in each of the ribbon feature blocks based on a chroma compensation evaluation value of each of the plurality of ribbon feature blocks After the value it also includes:
  • Each of the plurality of ribbon feature blocks is output to the ribbon compensation network in accordance with a target chrominance value for each of the pixel points.
  • the color distortion of the image to be processed is compensated by using a dithering algorithm according to the target chromaticity value, including:
  • the outputted plurality of ribbon feature blocks are superimposed with the image to be processed by a dithering algorithm to compensate for color distortion of the image to be processed.
  • the embodiment of the present disclosure further provides a color distortion compensating device, the device comprising:
  • a ribbon feature block obtaining module configured to input the image to be processed into the ribbon recognition neural network, and acquire a plurality of ribbon feature blocks
  • a chroma compensation evaluation value obtaining module configured to acquire the each according to a chroma value of each of the plurality of ribbon feature blocks and a chroma value of the adjacent ribbon feature block a chroma compensation evaluation value of the plurality of ribbon feature blocks; wherein the adjacent ribbon feature block is a ribbon feature block adjacent to each of the ribbon feature blocks;
  • a marking module configured to mark each of the ribbon feature blocks according to each of the ribbon feature block chromaticity compensation evaluation values
  • a target chrominance value determining module configured to input the marked plurality of ribbon feature blocks into the ribbon compensation network, and perform chromaticity compensation evaluation according to each of the plurality of ribbon feature blocks a value determining a target chrominance value for each pixel in each of the ribbon feature blocks;
  • the color distortion compensation module is configured to compensate for color distortion of the image to be processed by using a dithering algorithm according to the target chromaticity value.
  • the ribbon feature block acquisition module is configured to:
  • the convolution layer in the ribbon recognition neural network extracts a plurality of suspected ribbon blocks in the image to be processed; wherein the plurality of suspected ribbon regions The pixels in each of the suspected ribbon feature blocks are consecutive and the chrominance values of the pixels in each of the suspected ribbon feature blocks are the same;
  • the plurality of suspected ribbon blocks are analyzed using an activation function
  • Embodiments of the present disclosure also provide a television set including the above-described compensation device.
  • the ribbon feature block is identified by using the ribbon network, and then the target chromaticity value of the pixel in the ribbon feature block is determined by using the ribbon compensation network, and finally the color distortion compensation is implemented by using the dither function. It can effectively solve the band phenomenon caused by color distortion and improve the reliability of compensating for color distortion.
  • FIG. 1 is a flowchart of a method for compensating for color distortion according to an embodiment
  • FIG. 2 is a schematic structural diagram of a color distortion compensating apparatus according to an embodiment.
  • FIG. 1 is a flow chart of a method for compensating for color distortion provided by an embodiment.
  • This embodiment can solve the color band phenomenon caused by color distortion, and the method can be performed by a color distortion compensating device which can be integrated in an electronic product with video codec such as a digital television set.
  • the method includes the following steps.
  • Step 110 Input the image to be processed into the ribbon recognition neural network to acquire a plurality of ribbon feature blocks.
  • the ribbon recognition neural network can be obtained in a deep learning manner. For example, tens of thousands of pictures are trained as samples to search and identify the color bands in the sample picture, so that the neural network has the ability to accurately identify the color bands in the image.
  • the ribbon feature block may be an image region having a ribbon feature in the image to be processed.
  • the image to be processed is input into the ribbon recognition neural network
  • the process of acquiring the plurality of ribbon feature blocks may be: convolution in the ribbon recognition neural network after the image to be processed is input into the ribbon recognition neural network.
  • the layer extracts a plurality of suspected color band blocks in the image to be processed, wherein the pixel points in each of the plurality of suspected color band blocks are consecutive and the pixels in each suspected color band block The chroma values are the same.
  • the suspect function ribbon block is then analyzed using the activation function. If the plurality of suspected ribbon blocks satisfy the set condition, it is determined that the plurality of suspected ribbon blocks are a plurality of ribbon feature blocks.
  • the chrominance values between the plurality of suspected ribbon blocks may be the same or different.
  • the chrominance value can be a Red Green Blue (RGB) value.
  • the process of extracting a plurality of suspected color band blocks in the image to be processed may be: acquiring a chromaticity value of each pixel in the image to be processed, and analyzing the chromaticity values of all the pixels in the image to be processed, and continuously selecting the pixel points. The regions with the same chromaticity values are extracted as suspected ribbon blocks.
  • the expression for the activation function is: Where X is the input variable of the activation function.
  • the method for analyzing the plurality of suspected ribbon blocks by using the activation function may be: respectively acquiring the block sizes of the plurality of suspected color band blocks, and respectively determining the block sizes of the plurality of suspected color band blocks. Substituting the activation function for calculation, obtaining a plurality of first objective function values, and averaging the plurality of first objective function values to obtain a second objective function value.
  • the block size of a suspected ribbon block may be the number of pixels included in the suspected ribbon block.
  • the average value of the first objective function corresponding to the plurality of suspected ribbon blocks is averaged to obtain the second objective function value: Where N is the number of suspected ribbon blocks.
  • determining the plurality of suspected ribbon blocks into the plurality of ribbon feature blocks may be implemented by: if the second objective function value If it is less than the preset value, it is determined that the plurality of suspected ribbon blocks are a plurality of ribbon feature blocks.
  • the preset value is set to 0.9.
  • the second objective function value BL obtained in the above embodiment, if BL ⁇ 0.9, the plurality of suspected ribbon blocks are a plurality of ribbon feature blocks; if BL ⁇ 0.9, multiple suspects The ribbon block is not a ribbon feature block, that is, there is no ribbon phenomenon in the image to be processed.
  • Step 120 Acquire chromaticity compensation evaluation of each color ribbon feature block according to chrominance values of each of the plurality of ribbon feature blocks and chromaticity values of adjacent ribbon feature blocks value.
  • the adjacent ribbon feature block is a ribbon feature block adjacent to the current ribbon feature block, and the number of the ribbon feature blocks adjacent to the current ribbon feature block may be 1 or more.
  • the chromaticity of each color ribbon feature block is obtained according to the chromaticity value of each of the plurality of ribbon feature blocks and the chromaticity value of the adjacent ribbon feature block.
  • the process of compensating the evaluation value may be: calculating a difference between a chrominance value of each ribbon feature block and a chromaticity value of an adjacent ribbon feature block, and calculating an average value of the difference values to obtain each color band characteristic area.
  • the chroma compensation evaluation value of the block may be: calculating a difference between a chrominance value of each ribbon feature block and a chromaticity value of an adjacent ribbon feature block, and calculating an average value of the difference values to obtain each color band characteristic area.
  • calculating the difference between the chrominance value of each ribbon feature block and the chrominance value of the adjacent ribbon feature block may be respectively calculating the chromaticity value of each of the current ribbon feature blocks and each The difference in chrominance values for adjacent ribbon feature blocks.
  • the chrominance compensation is performed. The evaluation value is: Where M is the number of ribbon feature blocks adjacent to the current ribbon feature block.
  • Step 130 marking each of the ribbon feature blocks based on the chroma compensation evaluation value of each of the ribbon feature blocks.
  • the chrominance compensation evaluation values are marked in the corresponding ribbon feature blocks.
  • Step 140 Input the marked plurality of ribbon feature blocks into the ribbon compensation network, and determine each ribbon feature region according to the chromaticity compensation evaluation value of each of the plurality of ribbon feature blocks. The target chrominance value for each pixel in the block.
  • the ribbon compensation network may compensate for image chrominance based on the loss of chrominance in different regions of the image to be processed.
  • Ribbon Compensation Network and Ribbon Recognition Neural networks belong to different layers in a neural network and implement different functions.
  • the target chrominance value may be a chrominance value compensated for a pixel point in the ribbon feature block.
  • determining a target of each pixel in each ribbon feature block based on a chroma compensation evaluation value of a ribbon feature block of each of the plurality of ribbon feature blocks The chrominance value can be implemented by first determining the weight function of each ribbon feature block based on the chrominance compensation evaluation value of each of the plurality of ribbon feature blocks, and then based on each The weight function of the ribbon feature block calculates the weight value of each pixel in each ribbon feature block, and finally calculates each ribbon feature block according to the weight value of each pixel of each ribbon feature block.
  • the target chromaticity value for each pixel in the middle is the target chromaticity value for each pixel in the middle.
  • the weight function determined by the chrominance compensation evaluation value is: Where d(j,k) is the distance between the jth pixel point and the kth adjacent color band feature block in the current color band feature block, and I is the chromaticity value of the current color band characteristic block, I k For the chrominance value of the kth adjacent color band feature block, B 0 is the block size of the current color band feature block, and B k is the block size of the kth adjacent color band feature block.
  • represents the norm of I and I k . As can be seen from the weight function, when
  • the determining manner of d(j, k) may be the shortest value among the plurality of pixel points of the pixel point in the current ribbon feature block and the kth adjacent ribbon feature block.
  • the calculation formula of the target chromaticity value of each pixel is Where M is the number of ribbon feature blocks adjacent to the current ribbon feature block.
  • Step 150 using a dithering algorithm to compensate for color distortion of the image to be processed according to the target chromaticity value.
  • the dithering algorithm can solve the problem of looking at higher resolution images at low resolution, color and image distortion occurring when displaying more color modes in low color mode.
  • the method further includes: outputting the pixel points in the plurality of ribbon feature blocks to the ribbon compensation network according to the target chromaticity value.
  • the color distortion of the image to be processed is compensated according to the target chromaticity value by using a dithering algorithm, which may be implemented by using a dithering algorithm to superimpose the outputted plurality of ribbon feature blocks with the image to be processed. Compensate for color distortion of the image to be processed.
  • the technical solution of the embodiment uses the ribbon network to identify the ribbon feature block, and then uses the ribbon compensation network to determine the target chromaticity value of the pixel in the ribbon feature block, and finally uses the dither function to compensate the color distortion, which can be effective. It solves the band phenomenon caused by color distortion and improves the reliability of compensating for color distortion.
  • FIG. 2 is a schematic structural diagram of a color distortion compensating apparatus according to an embodiment. As shown in FIG. 2, the apparatus includes: a ribbon feature block acquisition module 210, a chrominance compensation evaluation value acquisition module 220, a labeling module 230, a target chrominance value determination module 240, and a color distortion compensation module 250.
  • the ribbon feature block obtaining module 210 is configured to input the image to be processed into the ribbon recognition neural network to acquire a plurality of ribbon feature blocks, and the chromaticity compensation evaluation value obtaining module 220 is configured to be configured according to the plurality of ribbon features Obtaining a chroma compensation value of each of the ribbon feature blocks and a chroma value of the adjacent ribbon feature block, and obtaining a chroma compensation evaluation value of each of the ribbon feature blocks; wherein the phase The adjacent ribbon feature block is a ribbon feature block adjacent to each of the ribbon feature blocks; and the marking module 230 is configured to perform an evaluation based on the chromaticity compensation value of each of the ribbon feature blocks Each of the ribbon feature blocks is marked; the target chrominance value determining module 240 is configured to input the marked plurality of ribbon feature blocks into the ribbon compensation network, according to the plurality of ribbon feature blocks The chrominance compensation evaluation value of each of the ribbon feature blocks determines a target chromaticity value of each pixel in each of the ribbon feature blocks; the
  • the ribbon feature block obtaining module 210 is configured to:
  • the convolution layer in the ribbon neural network extracts a plurality of suspected ribbon blocks in the image to be processed; wherein the plurality of suspected ribbon blocks Pixels in each of the suspected ribbon feature blocks are consecutive and the chroma values of the pixels in each of the suspected ribbon feature blocks are the same; the plurality of suspected ribbon blocks are performed using an activation function And analyzing; if the plurality of suspected ribbon blocks satisfy the setting condition, determining that the plurality of suspected ribbon blocks are a plurality of ribbon feature blocks.
  • the ribbon feature block obtaining module 210 is further configured to: respectively acquire the block sizes of the plurality of suspected color band blocks; and substitute the block size into an activation function to perform calculation a first objective function value; averaging the plurality of first objective function values to obtain a second objective function value; the ribbon feature block obtaining module 210 is configured to implement the plurality of suspects by: If the ribbon block satisfies the setting condition, determining that the plurality of suspected ribbon blocks are a plurality of ribbon feature blocks: if the second objective function value is less than a preset value, determining the plurality of suspect colors The strip block is a plurality of ribbon feature blocks.
  • the number of the ribbon feature blocks adjacent to each of the ribbon feature blocks is plural; the chroma compensation evaluation value obtaining module 220 is further configured to: calculate the a difference between a chrominance value of each of the ribbon feature blocks and a chromaticity value of the plurality of adjacent ribbon feature blocks; calculating an average of the plurality of the difference values to obtain the each of the ribbon feature blocks The chroma compensation evaluation value.
  • the target chrominance value determining module 240 is configured to: determine each of the ribbons according to chromaticity compensation evaluation values of each of the plurality of ribbon feature blocks a weight function of the feature block; calculating a weight value of each pixel of each of the ribbon feature blocks based on a weight function of each of the ribbon feature blocks; calculating each of the weight values according to the weight value The target chrominance value for each pixel in the ribbon feature block.
  • the method further includes: an output module configured to output pixel points in the plurality of ribbon feature blocks to the ribbon compensation network according to the target chromaticity value.
  • the color distortion compensation module 250 is further configured to: use the dithering algorithm to superimpose the output plurality of ribbon feature blocks with the image to be processed to compensate for color distortion of the image to be processed. .
  • the above apparatus can perform the methods provided by all the foregoing embodiments of the present disclosure, and has corresponding functional modules and advantageous effects for performing the above methods.
  • the above apparatus can perform the methods provided by all the foregoing embodiments of the present disclosure, and has corresponding functional modules and advantageous effects for performing the above methods.
  • the techniques not described in this embodiment reference may be made to the methods provided by all of the foregoing embodiments of the present disclosure.
  • the present disclosure further provides a television set including the above-mentioned color distortion compensating device.
  • the working principle of the color distortion compensating device in the television set can be referred to any of the above embodiments, and details are not described herein again.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Image Processing (AREA)
  • Facsimile Image Signal Circuits (AREA)
  • Color Image Communication Systems (AREA)
  • Image Analysis (AREA)

Abstract

La présente invention concerne un procédé et un appareil de compensation de distorsion des couleurs, et un téléviseur. Le procédé comporte les étapes consistant à: introduire une image à traiter dans un réseau neuronal de reconnaissance de ruban pour obtenir une pluralité de blocs de caractéristiques de ruban; acquérir une valeur d'évaluation de compensation de chrominance de chacun des blocs de caractéristiques de ruban selon la valeur de chrominance de chaque bloc de la pluralité de blocs de caractéristiques de ruban et la valeur de chrominance d'un bloc adjacent de caractéristiques de ruban; marquer chacun des blocs de caractéristiques de ruban selon la valeur d'évaluation de compensation de chrominance de chacun des blocs de caractéristiques de ruban; introduire la pluralité de blocs marqués de caractéristiques de ruban dans un réseau de compensation de ruban, et déterminer une valeur de chrominance cible de chaque pixel dans chacun des blocs de caractéristiques de ruban selon la valeur d'évaluation de compensation de chrominance de chaque bloc de la pluralité de blocs de caractéristiques de ruban; et compenser l'effet d'une distorsion des couleurs de l'image à traiter au moyen d'un algorithme de juxtaposition selon les valeurs de chrominance cibles.
PCT/CN2018/112335 2018-02-06 2018-10-29 Procédé et appareil de compensation de distorsion des couleurs, et téléviseur WO2019153799A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810116054.2 2018-02-06
CN201810116054.2A CN108322723B (zh) 2018-02-06 2018-02-06 一种色彩失真的补偿方法、装置和电视机

Publications (1)

Publication Number Publication Date
WO2019153799A1 true WO2019153799A1 (fr) 2019-08-15

Family

ID=62903566

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/112335 WO2019153799A1 (fr) 2018-02-06 2018-10-29 Procédé et appareil de compensation de distorsion des couleurs, et téléviseur

Country Status (2)

Country Link
CN (1) CN108322723B (fr)
WO (1) WO2019153799A1 (fr)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108322723B (zh) * 2018-02-06 2020-01-24 深圳创维－Rgb电子有限公司 一种色彩失真的补偿方法、装置和电视机
CN110930372B (zh) * 2019-11-06 2023-04-25 维沃移动通信有限公司 一种图像处理方法、电子设备及计算机可读存储介质

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1623165A (zh) * 2002-01-25 2005-06-01 汤姆森许可公司 用于降低轮廓的方法和系统
KR20050122782A (ko) * 2004-06-25 2005-12-29 엘지전자 주식회사 플라즈마 디스플레이 패널의 화상처리 방법
US20060233456A1 (en) * 2005-04-18 2006-10-19 Samsung Electronics Co., Ltd. Apparatus for removing false contour and method thereof
US20080123979A1 (en) * 2006-11-27 2008-05-29 Brian Schoner Method and system for digital image contour removal (dcr)
CN101583970A (zh) * 2007-01-19 2009-11-18 汤姆森许可贸易公司 减小数字图像中的轮廓
CN102568436A (zh) * 2010-12-16 2012-07-11 苹果公司 时空颜色亮度抖动技术
CN108027963A (zh) * 2015-09-14 2018-05-11 高通股份有限公司 抖动引导lut输出值内插
CN108322723A (zh) * 2018-02-06 2018-07-24 深圳创维－Rgb电子有限公司 一种色彩失真的补偿方法、装置和电视机

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1623165A (zh) * 2002-01-25 2005-06-01 汤姆森许可公司 用于降低轮廓的方法和系统
KR20050122782A (ko) * 2004-06-25 2005-12-29 엘지전자 주식회사 플라즈마 디스플레이 패널의 화상처리 방법
US20060233456A1 (en) * 2005-04-18 2006-10-19 Samsung Electronics Co., Ltd. Apparatus for removing false contour and method thereof
US20080123979A1 (en) * 2006-11-27 2008-05-29 Brian Schoner Method and system for digital image contour removal (dcr)
CN101583970A (zh) * 2007-01-19 2009-11-18 汤姆森许可贸易公司 减小数字图像中的轮廓
CN102568436A (zh) * 2010-12-16 2012-07-11 苹果公司 时空颜色亮度抖动技术
CN108027963A (zh) * 2015-09-14 2018-05-11 高通股份有限公司 抖动引导lut输出值内插
CN108322723A (zh) * 2018-02-06 2018-07-24 深圳创维－Rgb电子有限公司 一种色彩失真的补偿方法、装置和电视机

Also Published As

Publication number Publication date
CN108322723A (zh) 2018-07-24
CN108322723B (zh) 2020-01-24

Similar Documents

Publication Publication Date Title
Liu et al. CID: IQ–a new image quality database
WO2018161775A1 (fr) Procédé d'apprentissage de modèle de réseau neuronal, dispositif et support de stockage pour traitement d'image
Thakur et al. A new method for color image quality assessment
WO2016045242A1 (fr) Procédé d'agrandissement d'image, appareil d'agrandissement d'image et dispositif d'affichage
CN111489346B (zh) 一种全参考图像质量评价方法及系统
WO2016037422A1 (fr) Procédé de détection d'un changement d'une scène vidéo
US20180232862A1 (en) Image processing device and image processing method
CN109218716B (zh) 基于色彩统计和信息熵无参考色调映射图像质量评价方法
CN104504722B (zh) 一种利用灰色点校正图像颜色的方法
WO2017045213A1 (fr) Procédé et appareil de compensation de sous-pixel de panneau rvbb
WO2019153799A1 (fr) Procédé et appareil de compensation de distorsion des couleurs, et téléviseur
JP2013065215A (ja) 画像処理装置及び画像処理方法
Lee et al. Color image enhancement using histogram equalization method without changing hue and saturation
WO2019210707A1 (fr) Procédé d'évaluation de la netteté d'images, dispositif et dispositif électronique
CN102088539A (zh) 一种预拍照画质评价方法和系统
CN109658367A (zh) 基于彩色传递的图像融合方法
CN114202491B (zh) 一种增强光学图像的方法及系统
JP2010016803A (ja) 複数のカラーカメラ間の色調整装置および方法
CN101923455A (zh) 一种yuv格式的数字图像显示和分析方法
EP2672719A2 (fr) Étalonnage de la couleur d'un dispositif de capture d'images de manière à s'adapter à la scène à capturer
Hirao et al. Convolutional neural network based inverse tone mapping for high dynamic range display using LUCORE
RU2524869C1 (ru) Устройство колоризации черно-белых изображений
CN110298812A (zh) 一种图像融合处理的方法及装置
US20170337709A1 (en) Method to determine chromatic component of illumination sources of an image
Yuan et al. Color image quality assessment with multi deep convolutional networks

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: 18905164

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 03.12.2020)

122 Ep: pct application non-entry in european phase

Ref document number: 18905164

Country of ref document: EP

Kind code of ref document: A1