[go: up one dir, main page]

CN109118549A - A method of making object of reference with white printing paper and restores object color - Google Patents

A method of making object of reference with white printing paper and restores object color Download PDF

Info

Publication number
CN109118549A
CN109118549A CN201810804137.0A CN201810804137A CN109118549A CN 109118549 A CN109118549 A CN 109118549A CN 201810804137 A CN201810804137 A CN 201810804137A CN 109118549 A CN109118549 A CN 109118549A
Authority
CN
China
Prior art keywords
color
printing paper
white printing
neural network
artificial neural
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201810804137.0A
Other languages
Chinese (zh)
Inventor
贾振堂
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shanghai University of Electric Power
Original Assignee
Shanghai University of Electric Power
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 Shanghai University of Electric Power filed Critical Shanghai University of Electric Power
Priority to CN201810804137.0A priority Critical patent/CN109118549A/en
Publication of CN109118549A publication Critical patent/CN109118549A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/90—Determination of colour characteristics
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10004—Still image; Photographic image
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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 OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G06T2207/30088—Skin; Dermal

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Spectrometry And Color Measurement (AREA)
  • Image Processing (AREA)

Abstract

本发明涉及一种用白色打印纸作参照物恢复目标物颜色的方法，该方法包括如下步骤：(1)以白色打印纸作为参照物获取目标物照片；(2)获取目标物照片中的目标物平均色以及白色打印纸空白部分平均色；(3)将目标物平均色和白色打印纸空白部分平均色输入至预先训练的人工神经网络颜色恢复模型得到目标物颜色；所述的人工神经网络颜色恢复模型输入包括某一输入颜色在白色打印纸中的平均色以及白色打印纸空白部分平均色，人工神经网络颜色恢复模型输出为该输入颜色对应的标准色。与现有技术相比，本发明方法简便易行，颜色恢复结果准确可靠。

The invention relates to a method for recovering the color of a target object by using white printing paper as a reference object. The method includes the following steps: (1) taking a white printing paper as a reference object to obtain a photo of the target object; (2) obtaining the target object in the photo of the target object The average color of the object and the average color of the blank part of the white printing paper; (3) the average color of the target object and the average color of the blank part of the white printing paper are input into the pre-trained artificial neural network color recovery model to obtain the target color; the artificial neural network The input of the color recovery model includes the average color of a certain input color in the white printing paper and the average color of the blank part of the white printing paper, and the output of the artificial neural network color recovery model is the standard color corresponding to the input color. Compared with the prior art, the method of the invention is simple and easy to implement, and the color recovery result is accurate and reliable.

Description

A method of making object of reference with white printing paper and restores object color
Technical field
The present invention relates to image identifying and processing fields, make object of reference recovery mesh with white printing paper more particularly, to a kind of The method for marking object color.
Background technique
Same object shoots resulting photo under different illumination conditions can show different colors, make subsequent mesh The generations such as mark is other, the colour of skin calculates are difficult.For example, in the icterus neonatorum detection based on image procossing, different illumination conditions Lower newborn skin photo can generate difference, so that processing result image has differences, jaundice testing result is not accurate enough.Cause This, needs a kind of method that the color for acquiring resulting same object under different light environments is all restored to the standard of the target In color, to eliminate influence brought by photo environment difference.
Summary of the invention
It is an object of the present invention to overcome the above-mentioned drawbacks of the prior art and provide a kind of with white printing paper Make the method that object of reference restores object color.
The purpose of the present invention can be achieved through the following technical solutions:
A method of making object of reference with white printing paper and restore object color, this method comprises the following steps:
(1) object photo is obtained using white printing paper as object of reference；
(2) the object average color and white printing paper blank parts average color in object photo are obtained；
(3) object average color and white printing paper blank parts average color are input to artificial neural network trained in advance Network color Restoration model obtains object color；
The artificial neural network color Restoration model input includes that a certain input color is flat in white printing paper Homochromatic and white printing paper blank parts average color, the output of artificial neural network color Restoration model are that the input color is corresponding Reference colour.
Training obtains the artificial neural network color Restoration model as follows:
(11) select a set of color set as reference colour；
(12) multiple color is chosen in the color set and is printed on white printing paper；
(13) acquisition is printed with the photo of the white printing paper of color block, obtains training sample, and the training sample is defeated Enter the reference colour that data include the average color of each color block, printing paper blank parts average color and corresponding color block；
(14) artificial neural network is created, training sample is trained to obtain artificial neural network color Restoration model.
Step (12) specifically: the N kind color { c for the concentration that gets colorsi, i=1,2 ..., N }, ciIndicate i-th kind of color Reference colour prints on N kind color on white printing paper in the form of uniform color lump.
Step (13) specifically:
(13a) obtains the photo for being printed with the white printing paper of color block respectively under different light environments, and then obtains Obtain M photos；
(13b) obtains following data to any one photo: the average color c ' of the color block of i-th kind of color of acquisitioniWith And printing paper blank parts average color c ' in the photop；
(13c) constructs N number of training sample for the photo in step (13b), wherein i-th of training sample are as follows:
samplei={ inputi,outputi, inputi=(c 'i,c′p), outputi=c 'i, i=1,2 ..., N；
(13d) executes step (13b)~(13c) to M photos respectively, obtains M × N number of training sample.
Step (14) specifically: creation artificial neural network, the artificial neural network are averaged each color block Color and printing paper blank parts average color are as input, using the reference colour of corresponding color block as desired output, using instruction Practice sample artificial neural network is trained to obtain artificial neural network color Restoration model.
Step (1) specifically: by object and white printing paper close to putting together or then directly place object It takes pictures on white printing paper, keeps illumination consistent when taking pictures, the object photo of acquisition is reference with white printing paper Object.
Compared with prior art, the present invention has the advantage that
(1) present invention restores the method for object standard color by white printing paper, can will be in different light environments The color of the lower resulting same object of acquisition is all restored on the standard color of the target, to eliminate photo environment difference institute band The influence come；
(2) present invention requires with white printing paper to put object together and take pictures in training stage and application stage, And make them close as far as possible, the consistency of illumination is kept, target object area average color and printing paper region average color in photo Construct input sample, and the input as neural network, and using reference colour as the desired output of neural network, so that As a result more accurate and reliable；
(3) the method for the present invention by means of white paper out at, white printing paper is article very common in daily life, It is easy to obtain, and then this method is easy, it is easily operated.
Detailed description of the invention
Fig. 1 is the flow diagram that the present invention makees that object of reference restores the method for object color with white printing paper.
Specific embodiment
The present invention is described in detail with specific embodiment below in conjunction with the accompanying drawings.Note that the following embodiments and the accompanying drawings is said Bright is substantial illustration, and the present invention is not intended to be applicable in it object or its purposes is defined, and the present invention does not limit In the following embodiments and the accompanying drawings.
Embodiment
As shown in Figure 1, a kind of make the method that object of reference restores object color with white printing paper, this method includes as follows Step:
(1) object photo is obtained using white printing paper as object of reference；
(2) the object average color and white printing paper blank parts average color in object photo are obtained；
(3) object average color and white printing paper blank parts average color are input to artificial neural network trained in advance Network color Restoration model obtains object color；
The input of artificial neural network color Restoration model include a certain average color of the input color in white printing paper with And white printing paper blank parts average color, the output of artificial neural network color Restoration model are the corresponding standard of input color Color.
To sum up, the present invention includes two stages: training stage and application stage, and the training stage, i.e. training was artificial in advance Neural network color Restoration model, the application stage is i.e. using training artificial neural network color Restoration model in advance to recovery target The stage of object color.
One, the training stage
Training obtains artificial neural network color Restoration model as follows:
(11) select a set of color set as reference colour, present invention aim to the photo color of object is restored Onto some color in the color set.
(12) multiple color is chosen in the color set and is printed on white printing paper, and specifically: get colors concentration N kind color { ci, i=1,2 ..., N }, ciThe reference colour for indicating i-th kind of color beats N kind color in the form of uniform color lump It is printed on white printing paper, the size of N determines the accuracy for restoring color in the future, and N is bigger, and precision is higher.
(13) acquisition is printed with the photo of the white printing paper of color block, obtains training sample, training sample input data The reference colour of average color, printing paper blank parts average color and corresponding color block including each color block, specifically includes Following sub-step:
(13a) obtains the photo for being printed with the white printing paper of color block respectively under different light environments, and then obtains Obtain M photos；
(13b) obtains following data to any one photo: the average color c ' of the color block of i-th kind of color of acquisitioniWith And printing paper blank parts average color c ' in the photop；
(13c) constructs N number of training sample for the photo in step (13b), wherein i-th of training sample are as follows:
samplei={ inputi,outputi, inputi=(c 'i,c′p), outputi=c 'i, i=1,2 ..., N；
(13d) executes step (13b)~(13c) to M photos respectively, obtains M × N number of training sample.
(14) artificial neural network is created, training sample is trained to obtain artificial neural network color Restoration model, Specifically: creation artificial neural network, artificial neural network equal the average color of each color block and printing paper blank parts Equal color is as input, using the reference colour of corresponding color block as desired output, using training sample to artificial neural network into Row training obtains artificial neural network color Restoration model.
Two, the application stage
In use, by object and white printing paper close to putting together or object is then directly placed at white It takes pictures on printing paper, keeps illumination consistent when taking pictures, the object photo of acquisition is using white printing paper as reference substance.
Paper white area and target object area are extracted in the object photo taken, and calculates separately its average color, Wherein, object average color is c ', and white printing paper blank parts average color is c 'p, construct input sample input=(c ', c ′p), by input=(c ', c 'p) it is input to artificial neural network color Restoration model, it obtains artificial neural network color and restores mould The output output=c of type, c are the standard color to c ' reconstruction, also as object color.
As an embodiment, the method for recovery forearm skin color described herein to reference colour.In the example, use One five layers full connection artificial neural network, but the adoptable artificial neural network of the present invention is not limited to this.Using RGB Color space, but the present invention is not required for being confined to RGB color, the input of sample are as follows:And the input vector input of sample can also be defined as By aim colour c 'iWith blank sheet of paper color c 'pOther vector forms come are combined into, as long as in training and keeping one in two stages of application Cause.
Specifically:
Sample collection:
(1) in this example, reference colour set is using whole colors in RGB color.It is empty in the color of this cube Between 2500 points of middle sampling, obtain 2500 kinds of color { ci=(Ri,Gi,Bi), i=1,2 ..., N }, N=2500.
(2) according to these color values, 2500 uniform color segments is generated by computer, are arranged with 50x50, and pass through The higher printer of color fidelity prints to these color blocks on white printing paper, while reserving on white printing paper certain Blank parts.
(3) it under a variety of different light environments, takes pictures to the color block printed.To taking pictures to obtain any Digital photograph does following processing: the wherein average color of each color block and the average color of printing paper white space are acquired, Obtain { c 'i=(R 'i,G′i,B′i), i=1,2 ..., N }, c 'p=(R 'p,G′p,B′p)。
(4) it is based on above-mentioned average color, constructs N number of sample, samplei={ inputi,outputi, i=1,2 ..., N, In,outputi=ci=(Ri,Gi,Bi), superscript 2 For square.
(5) due to can shoot a large amount of photo in (3), such as M=100, therefore available a large amount of sample, such as MxN=100*2500=250000 sample.
Model training:
(6) artificial neural network is created, uses one 5 layers of full Connection Neural Network, the neuron of each layer here Quantity are as follows: first layer (i.e. input layer) 12, the second layer 100, third layer 300, the 4th layer 100, layer 5 (exports Layer) 3.Second and third, four layers below all follow one dropout layers, loss ratio rate is set as 0.25.
(7) artificial neural network is trained with above-mentioned 250000 samples, wherein inputiAs input, and outputiAs desired output.After training, an artificial neural network color Restoration model Model is obtained.
Model application:
(8) it in use, in order to restore the reference colour of small skin of arm (i.e. object) from digital photograph, needs taking pictures When one A4 white printing paper and forearm are placed among A4 white printing paper close to putting together, or forearm.Then right They take pictures, and make uniform illumination as far as possible.
(9) in this photo taken, paper white area and skin area are extracted, and calculates separately its average face Color obtains average color c '=(R ', G ', B ') of the small skin of arm and average color c ' of blank sheet of paperp=(R 'p,G′p,B′p).And Input sample is established according to this:
(10) input is input in artificial neural network color Restoration model Model, obtains it and exports output= (R,G,B).Here output is the reference colour rebuild for skin color c'.
Above embodiment is only to enumerate, and does not indicate limiting the scope of the invention.These embodiments can also be with other Various modes are implemented, and can make in the range of not departing from technical thought of the invention it is various omit, displacement, change.

Claims (6)

1.一种用白色打印纸作参照物恢复目标物颜色的方法，其特征在于，该方法包括如下步骤：1. a method of using white printing paper as a reference to restore the color of a target, is characterized in that, the method comprises the steps: (1)以白色打印纸作为参照物获取目标物照片；(1) Use white printing paper as a reference to obtain a photo of the target object; (2)获取目标物照片中的目标物平均色以及白色打印纸空白部分平均色；(2) Obtain the average color of the target object in the photo of the target object and the average color of the blank part of the white printing paper; (3)将目标物平均色和白色打印纸空白部分平均色输入至预先训练的人工神经网络颜色恢复模型得到目标物颜色；(3) Input the average color of the target object and the average color of the blank part of the white printing paper into the pre-trained artificial neural network color recovery model to obtain the color of the target object; 所述的人工神经网络颜色恢复模型输入包括某一输入颜色在白色打印纸中的平均色以及白色打印纸空白部分平均色，人工神经网络颜色恢复模型输出为该输入颜色对应的标准色。The input of the artificial neural network color recovery model includes the average color of a certain input color in the white printing paper and the average color of the blank part of the white printing paper, and the output of the artificial neural network color recovery model is the standard color corresponding to the input color. 2.根据权利要求1所述的一种用白色打印纸作参照物恢复目标物颜色的方法，其特征在于，所述的人工神经网络颜色恢复模型通过如下步骤训练得到：2. a kind of method that uses white printing paper as reference to restore target color according to claim 1, it is characterized in that, described artificial neural network color recovery model obtains by following steps training: (11)选择一套颜色集作为标准色；(11) Select a set of colors as the standard color; (12)在该颜色集中选取多种颜色并打印于白色打印纸上；(12) Select multiple colors in this color set and print on white printing paper; (13)采集打印有颜色块的白色打印纸的照片，获取训练样本，所述的训练样本输入数据包括各个颜色块的平均颜色、打印纸空白部分平均颜色和对应颜色块的标准色；(13) collect the photo of the white printing paper printed with the color blocks, obtain training samples, and the input data of the training samples include the average color of each color block, the average color of the blank part of the printing paper and the standard color of the corresponding color block; (14)创建人工神经网络，对训练样本进行训练得到人工神经网络颜色恢复模型。(14) Create an artificial neural network, and train the training samples to obtain an artificial neural network color restoration model. 3.根据权利要求2所述的一种用白色打印纸作参照物恢复目标物颜色的方法，其特征在于，步骤(12)具体为：选取颜色集中的N种颜色{ci，i＝1,2,…,N}，ci表示第i种颜色的标准色，将N种颜色以均匀色块的形式打印于白色打印纸上。3. a kind of method that uses white printing paper as a reference to restore the color of the target object according to claim 2, it is characterized in that, step (12) is specifically: select N kinds of colors {ci, i =1 in color set ,2,…,N}, c i represents the standard color of the ith color, and N colors are printed on white printing paper in the form of uniform color blocks. 4.根据权利要求3所述的一种用白色打印纸作参照物恢复目标物颜色的方法，其特征在于，步骤(13)具体为：4. a kind of method that uses white printing paper as reference object to restore the color of target object according to claim 3, it is characterized in that, step (13) is specifically: (13a)在不同的光照环境下分别获取打印有颜色块的白色打印纸的照片，进而获得M张照片；(13a) respectively obtain the photos of the white printing paper printed with color blocks under different illumination environments, and then obtain M photos; (13b)对任意一张照片获取如下数据：采集第i种颜色的颜色块的平均颜色c′i以及该照片中打印纸空白部分平均颜色c′p；(13b) Obtain the following data for any photo: collect the average color c′ i of the color block of the i-th color and the average color c′ p of the blank part of the printing paper in the photo; (13c)针对步骤(13b)中的照片构建N个训练样本，其中第i个训练样本为：(13c) Construct N training samples for the photos in step (13b), wherein the ith training sample is: samplei＝{inputi,outputi}，inputi＝(c′i,c′p)，outputi＝c′i，i＝1,2,…,N；sample i ={input i ,output i }, input i =(c' i ,c' p ), output i =c' i , i=1,2,...,N; (13d)对M张照片分别执行步骤(13b)～(13c)，得到M×N个训练样本。(13d) Steps (13b) to (13c) are respectively performed on the M photos to obtain M×N training samples. 5.根据权利要求2所述的一种用白色打印纸作参照物恢复目标物颜色的方法，其特征在于，步骤(14)具体为：创建人工神经网络，所述的人工神经网络将各个颜色块的平均颜色和打印纸空白部分平均颜色作为输入，将对应颜色块的标准色作为期望输出，采用训练样本对人工神经网络进行训练得到人工神经网络颜色恢复模型。5. a kind of method that uses white printing paper as a reference to restore the color of the target object according to claim 2, it is characterized in that, step (14) is specifically: create artificial neural network, described artificial neural network will each color The average color of the block and the average color of the blank part of the printing paper are used as input, the standard color of the corresponding color block is used as the expected output, and the artificial neural network is trained by training samples to obtain the artificial neural network color restoration model. 6.根据权利要求1所述的一种用白色打印纸作参照物恢复目标物颜色的方法，其特征在于，步骤(1)具体为：将目标物与白色打印纸紧邻放在一起或者将目标物则直接放置在白色打印纸上进行拍照，拍照时保持光照一致，获得的目标物照片以白色打印纸为参考物。6. The method for recovering the color of a target object with white printing paper as a reference object according to claim 1, wherein the step (1) is specifically: placing the target object and the white printing paper next to each other or placing the target object The object is directly placed on the white printing paper to take pictures, and the illumination is kept consistent when taking pictures.
CN201810804137.0A 2018-07-20 2018-07-20 A method of making object of reference with white printing paper and restores object color Pending CN109118549A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810804137.0A CN109118549A (en) 2018-07-20 2018-07-20 A method of making object of reference with white printing paper and restores object color

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810804137.0A CN109118549A (en) 2018-07-20 2018-07-20 A method of making object of reference with white printing paper and restores object color

Publications (1)

Publication Number Publication Date
CN109118549A true CN109118549A (en) 2019-01-01

Family

ID=64862909

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810804137.0A Pending CN109118549A (en) 2018-07-20 2018-07-20 A method of making object of reference with white printing paper and restores object color

Country Status (1)

Country Link
CN (1) CN109118549A (en)

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101039439A (en) * 2007-04-23 2007-09-19 北京中星微电子有限公司 Method and apparatus for realizing correction of white balance
CN101706964A (en) * 2009-08-27 2010-05-12 北京交通大学 Color constancy calculating method and system based on derivative structure of image
CN103839236A (en) * 2014-02-25 2014-06-04 中国科学院自动化研究所 Image white balance method based on sparse representation
CN104572538A (en) * 2014-12-31 2015-04-29 北京工业大学 K-PLS regression model based traditional Chinese medicine tongue image color correction method
CN105738292A (en) * 2014-12-12 2016-07-06 中国科学院大连化学物理研究所 Cell phone imaging method with colorimetric sensor array
CN106468661A (en) * 2016-09-12 2017-03-01 深圳市飞点健康管理有限公司 A kind of urine test paper colour imaging correlation detection methods and system
CN106683140A (en) * 2016-12-16 2017-05-17 深圳市中达瑞和科技有限公司 Color recognition method and system
CN107507250A (en) * 2017-06-02 2017-12-22 北京工业大学 A kind of complexion tongue color image color correction method based on convolutional neural networks
CN107578390A (en) * 2017-09-14 2018-01-12 长沙全度影像科技有限公司 A kind of method and device that image white balance correction is carried out using neutral net
CN108205671A (en) * 2016-12-16 2018-06-26 浙江宇视科技有限公司 Image processing method and device

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101039439A (en) * 2007-04-23 2007-09-19 北京中星微电子有限公司 Method and apparatus for realizing correction of white balance
CN101706964A (en) * 2009-08-27 2010-05-12 北京交通大学 Color constancy calculating method and system based on derivative structure of image
CN103839236A (en) * 2014-02-25 2014-06-04 中国科学院自动化研究所 Image white balance method based on sparse representation
CN105738292A (en) * 2014-12-12 2016-07-06 中国科学院大连化学物理研究所 Cell phone imaging method with colorimetric sensor array
CN104572538A (en) * 2014-12-31 2015-04-29 北京工业大学 K-PLS regression model based traditional Chinese medicine tongue image color correction method
CN106468661A (en) * 2016-09-12 2017-03-01 深圳市飞点健康管理有限公司 A kind of urine test paper colour imaging correlation detection methods and system
CN106683140A (en) * 2016-12-16 2017-05-17 深圳市中达瑞和科技有限公司 Color recognition method and system
CN108205671A (en) * 2016-12-16 2018-06-26 浙江宇视科技有限公司 Image processing method and device
CN107507250A (en) * 2017-06-02 2017-12-22 北京工业大学 A kind of complexion tongue color image color correction method based on convolutional neural networks
CN107578390A (en) * 2017-09-14 2018-01-12 长沙全度影像科技有限公司 A kind of method and device that image white balance correction is carried out using neutral net

Similar Documents

Publication Publication Date Title
CN107507250B (en) Surface color and tongue color image color correction method based on convolutional neural network
Zhou et al. Learning data-driven reflectance priors for intrinsic image decomposition
Bychkovsky et al. Learning photographic global tonal adjustment with a database of input/output image pairs
Palermo et al. Dating historical color images
CN109086826A (en) Wheat Drought recognition methods based on picture depth study
CN111460947B (en) BP neural network-based method and system for identifying metal minerals under microscope
CN109977780A (en) A kind of detection and recognition methods of the diatom based on deep learning algorithm
CN111562010B (en) A method and apparatus for automated image color calibration
US10013784B2 (en) Generating an assembled group image from subject images
CN113269191A (en) Crop leaf disease identification method and device and storage medium
JP2013242825A (en) Personal attribute estimation system and learning data creation device
CN101966083A (en) Abnormal skin area calculation system and calculation method thereof
CN110074788A (en) A kind of body data acquisition methods and device based on machine learning
CN103839236A (en) Image white balance method based on sparse representation
CN114708519B (en) Elk identification and morphological contour parameter extraction method based on unmanned aerial vehicle remote sensing
Amirshahi et al. Judging aesthetic quality in paintings based on artistic inspired color features
CN116310549A (en) A detection method for invasive plant seeds in imported soybeans
US12482179B2 (en) Freestyle acquisition method for high-dimensional material
CN109410158B (en) Multi-focus image fusion method based on convolutional neural network
CN113511028A (en) Oil painting copying method
CN109118549A (en) A method of making object of reference with white printing paper and restores object color
US20030012435A1 (en) Apparatus and method for machine vision
CN106289089A (en) The optimization preparation method of experiment speckle field
CN114494038B (en) Target surface perspective distortion correction method based on improved YOLOX-S
CN110706296A (en) Batch automatic detection method for background color compliance of electronic certificate photos

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20190101

RJ01 Rejection of invention patent application after publication