CN111429547B - A Synthetic Method of Abnormal Color Vision Test Chart Based on False Same Color Search - Google Patents

Abstract

The invention relates to a method for synthesizing abnormal color vision test charts based on false same-color search. The main steps include: searching for false same-color pairs; selecting false same-color pairs; generating test images; the present invention is aimed at testing the color vision of subjects in abnormal color vision tests With the help of the abnormal vision simulation algorithm, the pseudo-same color pairs in the color space are selected, and binary color filling is performed on the abnormal vision test pattern, so as to generate abnormalities with diverse content, random repeatability, and specificity for abnormal color vision types Compared with the existing manual drawing or computer-aided semi-manual drawing method, the color vision test chart is flexible, random, simple and controllable, and has specificity for different abnormal color vision.

Description

A Synthetic Method of Abnormal Color Vision Test Chart Based on False Same Color Search

technical field

The invention belongs to the technical field of digital image generation, and in particular relates to a method for synthesizing abnormal color vision test charts.

Background technique

With the development of social economy and the continuous advancement of science and technology, the division of professional technology is becoming more and more refined, people are more concerned about the health status of individual eyesight and vision, and many professions have higher requirements for color discrimination ability. Some industries and departments have more specific requirements for the eyesight and vision of their employees. For example, there are clear regulations on vision and color vision in the fields of air force pilot medical examination and military enlistment medical examination. The application of color-blind students has also been restricted to a certain extent. Foreign countries have strict requirements on the visual and color vision of some special service units. As a widely recognized scientific quantitative inspection method, visual acuity test is particularly important for the discovery of talents, the field of health and even the development of social economy. Among them, the color-blind color vision test chart based on the false same color principle is the most commonly used color vision test tool in the world.

Color blindness color vision test chart is a key tool for detecting color blindness. The inspection method designed according to the principle of pseudo-same color has been used for screening abnormal color vision for a hundred years. Through various types of inspection patterns, such as geometric figures, digital figures, line figures, object figures, etc., it can detect abnormalities of red, green, blue and yellow and other color vision. The main disadvantages of this method of color blindness inspection using false synchromatic principle atlas are single pattern, fixed color scheme, non-specificity, etc. These factors will seriously affect the objective authenticity of the test results.

Most of the existing abnormal color vision test charts are inspection picture books with a fixed style drawn by hand or with computer drawing tools such as Photoshop. The style is single, the paradigm is fixed, and the drawing is relatively cumbersome. For example, the fifth edition of the "Color Blindness Test Chart" widely used in medical examinations, the first set of charts can be used for large-scale rapid examination; the second set of charts is characterized by simple geometric figures, suitable for adults with low education and illiterate physical examination The third group of pictures is suitable for checking children, the fourth group of pictures is a multi-digit group, which is used for physical examination of professionals who have high requirements for color vision; the fifth group of pictures is acquired color vision examination charts, suitable for clinical ophthalmologists, Auxiliary diagnosis for fundus diseases and central nervous system diseases by neurologists and surgeons.

The principle of color vision in the abnormal color vision test chart is that light red and light green intersect to interfere with the tester. If he is color blind, he can only see one color, so he cannot distinguish the numbers formed by the color difference. The general color tone of the graph of the correct answer in the picture is different from the background color. During the inspection process, medical staff often use these existing artificial color blindness inspection charts to test patients one by one. Rougher conditions. In particular, the picture book has a single pattern, and the testers can often use the method of association or marking memory to remember the answer without careful identification, which affects the accuracy and scientificity of the test results to a certain extent. The original manual design and computer-aided drawing methods have an impact on the accuracy and practicability of the abnormal color vision test chart due to the complex process, time-consuming and labor-intensive, single fixed style, and weak randomness. Therefore, it is necessary to study new methods. Method to make abnormal color vision test chart.

In addition, most of the existing computer-related color blindness instruments store the fixed patterns drawn in the existing color blindness test charts, and cannot flexibly and randomly generate new patterns according to the actual situation. For example, Shenzhen Luojie Technology Co., Ltd. provides a color blindness instrument for detecting color blindness in the patent "A Color Blindness Instrument and Color Blindness Detection Method" (patent application number: 20110224333 publication number: 102283632A). This detection method consists of two stages, the first stage is mainly to randomly display 21 fixed color blindness detection pictures to test the testees, its shortcoming is still relying on fixed detection patterns, randomly selected, but not randomly generated , cannot fundamentally solve the single image and cannot guarantee the objectivity of the detection results.

Contents of the invention

In view of the above technical problems, the present invention provides a method for synthesizing abnormal color vision test charts based on false same-color search, including:

Step 1: Search for false same-color pairs, that is, traverse and simulate the candidate colors on the RGB grid space, and calculate the perceived color difference before and after the simulation of the candidate colors;

Step 2: Select false same-color pairs, that is, first judge the threshold condition, then select false same-color pairs, and finally store false same-color pairs;

Step 3: Generate a test image, that is, select a color blindness test pattern, fill the foreground and background with false same color pairs, form a unit test pattern, and group the unit test patterns.

Further, step 1 includes:

Step 1.1: Select candidate colors, that is, divide the RGB space into grids, and randomly perform small offsets on the nodes of the grid or with each grid node as the center to select candidate colors;

通过色盲仿真算法得到其对应色向量/>

其中τ∈[d,p,t]，列表项[d,p,t]分别代表绿色盲、红色盲、蓝色盲；Step 1.2: Calculate the simulated color, that is, the selected triplet candidate color vector

Obtain its corresponding color vector through the color blindness simulation algorithm />

Among them, τ∈[d,p,t], and the list items [d,p,t] represent deuteranopia, protanopia, and tritanopia respectively;

Step 1.3: Calculate the perceptual color difference, that is, use the CIE L*a*b*1976 color difference formula to first convert the RGB color to the XYZ color space, then convert to the LAB color space, and then calculate {Q,Q ^τ } between Color difference value ΔE* _Lab (Q,Q ^τ );

Step 2 includes:

Step 2.1: Judging the threshold condition, that is, setting the threshold condition as α≥0, when ΔE* _Lab (Q,Q ^τ )≥α, keep the false same-color pair {Q,Q ^τ } as the element in the filling color set; When ΔE* _Lab (Q,Q ^τ )≤α, discard the false same color pair {Q,Q ^τ }, where α is a parameter that can be set, and its physical meaning is the color difference threshold;

Step 2.2: Select false same-color pairs, that is, calculate all candidate colors and judge the above threshold conditions, and select elements in the false same-color pair set according to the set color difference threshold α;

Step 2.3: Store false same-color pairs, that is, store the selected false same-color sets;

Step 3 includes:

Step 3.1: Select a test pattern, that is, select an easy-to-recognize geometric shape and a common figure in daily life as the test pattern, combine any false same color pair generated above as the foreground color and background color, and fill the foreground and background of the test pattern ;

Step 3.2: filling the test pattern, that is, using the selected false same color pair to fill the test pattern;

Step 3.3: Generate a test unit, that is, after filling the test pattern, a three-channel color image is obtained as a test unit;

Step 3.4: Group test images, that is, randomly select multiple test units and combine them into a set of test images.

Further, in step 1.1, divide the RGB color space into a grid of X×Y×Z, the size of which is 32×32×32, and select triplet candidate color vectors on the nodes of the grid;

The calculation of the corresponding color vector Q ^τ in step 1.2 is as follows: first, the triplet candidate color vector Q is converted to the LMS space, and the corresponding color vector in the LMS space is obtained, as follows:

为RGB色彩空间到LMS色彩空间的变换矩阵；in

is the transformation matrix from RGB color space to LMS color space;

Secondly, use the color-blindness simulation algorithm to obtain the τ-type color-blindness simulation vector in the LMS color space:

Where T _sim ^τ is the transformation matrix in the color blindness simulation algorithm;

Then convert the τ-type color blindness simulation vector of LMS color space back to RGB space vector, that is,

Among them, Q ^τ is the color-blind simulation corresponding color of the triplet candidate color vector Q, that is, {Q, Q ^τ } is a pair of false same colors;

In step 1.3, calculate the color difference value between {Q, Q ^τ } according to the following formula:

表示仿真色彩Q^τ转换到LAB空间后的三个分量；Where (L _Q *, a _Q *, b _Q *) represents the result of the candidate color Q converted to the LAB space,

Represents the three components after the simulation color Q ^τ is converted to the LAB space;

Adopt α=50 in step 2.1;

Step 2.2 randomly selects from the selected set of false same-color pairs, or directly picks up the color elements in the set after visualizing them;

The storage method in step 2.3 can adopt data structures such as hash list and dictionary, and the file format can use JSON or Pickle object;

In step 3.1, the number 0-9 is randomly generated as the foreground content, and the remaining blank area is used as the background to form a test pattern;

In step 3.4, select 5*5 unit test images to form a test image.

Further, in step 1.1, the RGB color space is divided into X×Y×Z grids, with the grid node (x ₀ , y ₀ , z ₀ ) as the center, and a random small step offset (Δx ,Δy,Δz) to reach the point, as a triplet candidate color vector.

Further, in step 3.2, the test pattern is directly filled with the false same color pair, the simulated color is used as the background filling color, and the original color is used as the foreground filling color.

Further, in step 3.2, the false same-color pairs are used to fill the split test pattern, and the K-D tree data structure is used to draw geometric patterns of small circles that do not overlap each other to form a split test pattern unit.

Further, in step 3.2, the approximate color of the false same color pair is used to fill the test pattern in a generalized manner, and in the LAB color space, the false same color is used as the center of the sphere and the radius is ρ to randomly pick similar colors according to the Gaussian distribution to fill Small geometric patterns generated each time, where 0<ρ<32.

其中Δx,Δy,Δz均为整数；Further, step 1.1 takes the modulus of the small step size

Where Δx, Δy, Δz are all integers;

In step 3.2, fill the test pattern with a fine-density checkered pattern.

The present invention also provides an application device for an abnormal color vision test chart, which is characterized in that: the image generated by the above-mentioned abnormal color vision test chart production method can be used in a special color vision detection computer device, embedded software and hardware, paper printing device or Internet application software system.

Further, the computer device for color vision detection includes various types of computers, mobile terminals, virtual reality equipment, Google glasses, iPad tablet computers, special displays, medical optometry testers and other equipment to develop corresponding offline or online applications in the software system to realize the above color vision The synthesis method of the test pattern, taking the unit test pattern and grouping test pattern as examples, presents on the display screen; the embedded software and hardware and the paper printing device are implemented in the corresponding programming language through various embedded systems such as single-chip microcomputers and FPGAs and development boards Color vision test chart synthesis method, or pre-synthesize color vision test charts on other hardware platforms, and then store them in embedded devices for display or printing; Internet application software systems are developed and implemented with various programming languages such as Javascript, hypertext markup language, etc. The method for synthesizing the color vision test is to present the color vision test diagram in the browser or APP in the form of micro-services and offline applications and implement the color vision test in combination with interactive design.

The present invention aims at the test pattern used to test the color vision of the subject in the abnormal color vision test, by means of the abnormal vision simulation algorithm, selects the false same color pair in the color space, and performs binary color filling on the abnormal vision test pattern, thereby producing Abnormal color vision test charts with diverse content, random repeatability, and specificity for abnormal vision types. Compared with the existing manual drawing or computer-aided semi-manual drawing method, the present invention has the characteristics of being flexible and random, simple and controllable, specific for different abnormal color visions, and the like.

Description of drawings

Fig. 1 is a flow chart of the method of the present invention;

Figure 2 shows the distribution curves of false homochromatic pairs corresponding to three typical types of color blindness in the RGB color space under different color thresholds α;

Fig. 3 is the unit test figure that uses the inventive method to make;

Fig. 4 is a grouping test diagram made using the method of the present invention.

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

In view of the shortcomings of the existing color vision inspection technology, as well as the problems of the single pattern of the existing abnormal color vision test chart and the non-specific color matching scheme, the present invention proposes a method for making an abnormal color vision test chart based on false same-color search. The technical idea is to use the existing abnormal color vision simulation algorithm to set parameters in the perceptual color space to search for the false same-color pair combination for a specific type of abnormal color vision; then select the pattern and use the false same-color pair to fill the foreground and background colors , to generate a binarized unit test chart; finally, multiple binarized unit test charts are grouped to form a test chart.

Concrete steps of the present invention are as shown in Figure 1, comprise:

Step 1: Search for false same-color pairs. Traverse and simulate the candidate colors on the RGB grid space, and calculate the perceived color difference before and after the simulation of the candidate colors.

Step 1.1: Select candidate colors. The RGB space is divided into grids, and a small offset is randomly performed on the nodes of the grid or centered on each grid node to select candidate colors;

其中Δx,Δy,Δz均为整数。Divide the RGB color space into X×Y×Z grids, for example, the size can be 32×32×32, and select triplet candidate color vectors on grid nodes; or use grid nodes (x ₀ ,y ₀ , z ₀ ) as the center, and shift (Δx, Δy, Δz) to each direction with a random small step size to reach the point, as a triplet candidate color vector, the modulus of the small step size

Among them, Δx, Δy, and Δz are all integers.

通过色盲仿真算法得到其对应色向量/>

其中τ∈[d,p,t]，列表项[d,p,t]分别代表绿色盲、红色盲、蓝色盲，对应色向量Q^τ的计算如下：Step 1.2: Calculate the simulated color. For the selected triplet candidate color vector

Obtain its corresponding color vector through color blindness simulation algorithm />

Among them, τ∈[d,p,t], the list items [d,p,t] represent deuteranopia, protanopia, and tritanopia respectively, and the corresponding color vector Q ^τ is calculated as follows:

First, the triplet candidate color vector Q is converted to the LMS space, and the corresponding color vector in the LMS space is obtained, as follows:

为RGB色彩空间到LMS色彩空间的变换矩阵；in

is the transformation matrix from RGB color space to LMS color space;

Secondly, use the color-blindness simulation algorithm proposed by Brettel et al. to obtain the τ-type color-blindness simulation vector in the LMS color space:

Where T _sim ^τ is the transformation matrix in the color blindness simulation algorithm;

Then convert the τ-type color blindness simulation vector of LMS color space back to RGB space vector, that is,

The above-mentioned series of calculation processes can be summarized as a single functional relational expression, namely:

Q ^τ = f _sim (Q; τ), τ∈[d,p,t]

Among them, Q ^τ is the color-blind simulation corresponding color of the triplet candidate color vector Q, that is, {Q, Q ^τ } is a pair of false same colors;

Step 1.3: Calculate the perceived color difference. Using the CIE L*a*b*1976 color difference formula, the RGB color is first converted to the XYZ color space, and then converted to the LAB color space, and then the color difference value between {Q,Q ^τ } is calculated according to the following formula:

表示仿真色彩Q^τ转换到LAB空间后的三个分量；Where (L _Q *, a _Q *, b _Q *) represents the result of the candidate color Q converted to the LAB space,

Represents the three components after the simulation color Q ^τ is converted to the LAB space;

Step 2: Select false same color pair. The threshold condition is judged first, then the false same-color pair is selected, and finally the false same-color pair is stored.

Step 2.1: Determine the threshold condition. Set the threshold condition as α≥0, when ΔE* _Lab (Q,Q ^τ )≥α, keep the false same color pair {Q,Q ^τ } as the element in the fill color set; when ΔE* _Lab (Q, When Q ^τ )≤α, discard the false homochromatic pair {Q,Q ^τ }; where α is a parameter that can be set, and its physical meaning is the threshold value of color difference. =50, at this time, the false homochromatic proportion of the red and green blind types is about 20%, and the false homochromatic proportion of the tritanopic type is about 70%. A sufficient number of false same colors can be selected for each type of color blindness;

Step 2.2: Select false same-color pairs. Calculate all candidate colors and judge the above threshold conditions, and select the elements in the false same color pair set according to the set color difference threshold α. In the actual filling process, you can randomly select from the selected set of false same-color pairs, or directly pick up the color elements in the set after visualization, so as to facilitate the selection of false-same-color pairs to be used;

Step 2.3: Store false same-color pairs. Store the selected false same-color set to avoid repeated calculation each time and facilitate subsequent use. The storage method can adopt data structures such as hash list and dictionary, and the file format can use JSON or Pickle object.

Step 3: Generate test images. Select the pattern of the color blindness test chart, and fill the foreground and background with false same color pairs to form a unit test chart; you can also group the unit test charts to form a group test chart.

Step 3.1: Select the test pattern. The selection of test patterns can be adjusted according to the actual situation, and geometric shapes that are easy to identify and common graphics in daily life can be used to reduce the interference of patterns on test results. Taking the Arabic numerals 0-9 as an example, the numbers 0-9 are randomly generated as the foreground content, and the remaining blank area is used as the background to form a test pattern. Combining any of the false same color pairs generated above as the foreground color and background color respectively, filling the foreground and background of the test pattern, a test image unit can be generated, and the generated test image unit at this time is a binary image, namely Contains only two colors, and these two colors contain only one color in the eyes of the corresponding color-blind person, so they cannot recognize the numbers in it;

Step 3.2: Fill the test pattern. Fill the test pattern with the selected pseudo-identical color pairs.

Mode 1: directly fill the test pattern with false same-color pairs to form a binary test pattern unit, as shown in Figure 3(a). Fill the foreground and background of the test pattern with the false same-color pairs selected in the above-mentioned false same-color search process. The resulting image, in the eyes of a person with abnormal color vision of the corresponding type, is a figure of the same color, so the foreground content cannot be recognized;

Mode 2: fill the test pattern with false same-color pairs to form a split test pattern unit, as shown in Figure 3(b). The false same-color pairs selected in the above-mentioned false same-color search process are respectively the foreground and the background, and the order can be arbitrary. The test pattern is used as the base plate to draw geometric patterns such as small circles that do not overlap each other. The size of the overlapping area between the geometric pattern area and the base plate , to determine whether the pattern belongs to the foreground or the background, so as to fill the corresponding color. In the drawing process of geometric patterns such as small circles that do not overlap with each other, various existing K-D tree data structures can be used to store attributes and coordinates of the drawn small geometric patterns to speed up the drawing process. The resulting color vision test chart is similar in texture to the color vision test chart commonly used in clinical practice. From the perspective of people with abnormal color vision of the corresponding type, the foreground is confused with the background, so the foreground content cannot be recognized;

Method 3: The approximate color generalization filling test pattern of the false same color pair generates a test unit with certain robustness, as shown in Fig. 3(c). Using the divided filling method described in method 2, when filling geometric patterns such as small circles, the false same color pair is not directly used, but in the sphere space with the false same color as the center and radius ρ in the LAB color space Randomly pick similar colors according to the Gaussian distribution to fill the small geometric patterns generated each time, generally 0<ρ<32. The resulting image is immune to interference from poorly calibrated displays, printers, and non-ideal brightness environments. In the eyes of people with abnormal color vision of the corresponding type, the foreground and background pattern textures are consistent, so the foreground content cannot be recognized;

Other ways: Other derivatives similar to the above three forms can also be included, such as a fine-density checkered pattern, or multiple foreground patterns presented side by side in each test unit. The resulting color vision test chart unit, in the eyes of people with abnormal color vision, will confuse the foreground with the background, and it is difficult to recognize the foreground content;

Step 3.3: Generate test units. After filling the test pattern, a three-channel color image can be obtained as a test unit. In addition, the candidate test unit can also be simulated to visually evaluate the effect of the image test unit, and further choose it according to whether it can theoretically confuse abnormal color vision.

Step 3.4: Group test images. Randomly select multiple test units and merge them into a set of test images, as shown in Figure 4, where Figure 4 (a), (b), and (c) respectively correspond to the test units that are filled in three ways in step 3.2. Group test graphs.

Considering the possible errors in the color blindness simulation model and the subtle differences in individual color vision objectively, multiple images are grouped together to form a test image, so that the number of test units read out by the subjects accounts for the overall test The ratio of the number of units to comprehensively evaluate the color vision of the subjects. Generally, 5*5 unit test diagrams can be selected to form a test image, and the unit test diagrams in the whole example diagram are all re-generated randomly during each test, thus ensuring the randomness of the test diagram.

The present invention also includes the following software and hardware devices and systems:

1. A computer device for detecting color vision. Develop corresponding offline or online applications for various types of computers, mobile terminals, virtual reality (VR) devices such as Oculus, Google Glass, iPad tablet computers, special displays, medical ophthalmology testers, etc., and realize the above color vision test in the software system The graph synthesis method uses unit test graphs and group test graphs as examples, and presents them on the display screen; it can also generate test graph samples described in this method in advance, store them in the hard disk or memory, and randomly select them for display during testing; at the same time, combined with special A series of evaluation procedures have been formulated to conduct color vision tests, including but not limited to scoring the results of the color vision test based on the correct reading rate of the grouped test charts. The input of control commands for image generation and presentation can be in various forms such as touch, voice, and mouse.

2. Embedded hardware and software and paper printing device. Through various embedded systems such as single-chip microcomputers and FPGAs and development boards, the method of synthesizing color vision test charts is realized in corresponding programming languages, or the color vision test charts are pre-synthesized on other hardware platforms, and then stored in embedded devices for display or printing. The color vision test chart is printed into a paper version in color, combined with the explanatory text of the test procedure, and distributed to the subjects in the form of test papers for testing. The printing equipment can adopt various models of color printing devices produced by various manufacturers at home and abroad, and can also adopt equipment such as rapid imaging camera technology and instant cameras. Specific implementation methods include but are not limited to the following methods: such as using embedded hardware and computer software to realize the production of color vision test charts, transmit them to color printers for printing, and distribute them to test personnel for color vision testing.

3. Internet application software system. Develop and implement the color vision test synthesis method with various programming languages such as Javascript, hypertext markup language, etc., or present the color vision test chart in the browser or APP in the form of micro-services and offline applications, which can be combined with certain interactive design To carry out the color vision test. The specific implementation architecture includes but is not limited to the following two methods: 1. Directly use Javascript language to execute in the client browser to generate and test the color vision test chart; 2. Use the background system to generate the color vision test chart, and then run it on the client browser Visually display in HTML or Single Page Application components for color vision testing.

The present invention can randomly generate pattern content on demand, which overcomes the defects of fixed patterns; parameters can be adjusted during the generation process of the present invention, and has the advantages of controllable color matching, scientific and easy-to-use; the present invention can be qualitatively judged by a single test chart For the color vision of the subjects, the severity of their color blindness can also be evaluated linearly according to the ratio of the subjects' recognition of the grouped unit test charts.

Claims (9)

1. An abnormal color vision test chart synthesis method based on pseudo-same color search comprises the following steps:

step 1: searching pseudo-same color pairs, namely traversing and simulating candidate colors on an RGB grid space, and calculating perceived color difference before and after the candidate color simulation;

the step 1 comprises the following steps:

step 1.1: selecting candidate colors, namely performing grid division on an RGB space, and randomly performing tiny offset on nodes of grids or taking each grid node as a center to select the candidate colors;

step 1.2: calculating the artificial colour, i.e. candidate colour vector for selected triples

The corresponding color vector +.A corresponding color vector is obtained through a color blind simulation algorithm>

Wherein τ ε [ d, p, t ]]List items [ d, p, t ]]Respectively representing green color blindness, red color blindness and blue color blindness;

step 1.3: calculating perceived color difference, i.e. converting RGB color to XYZ color space, then to LAB color space, and then calculating { Q, Q }, using CIE Lab1976 color difference formula ^τ Color difference value Δe between } _Lab (Q,Q ^τ )；

Step 2: selecting a false same-color pair, namely judging a threshold condition, selecting the false same-color pair, and storing the false same-color pair;

the step 2 comprises the following steps:

step 2.1: judging the threshold condition, namely setting the threshold condition to be alpha more than or equal to 0, when delta E is _Lab (Q,Q ^τ ) When not less than alpha, the pseudo-homochromatic pair { Q, Q is reserved ^τ -as an element in the set of fill colors; when delta E _Lab (Q,Q ^τ ) Discarding pseudo-homochromatic pairs when alpha is less than or equal to alpha

{Q,Q ^τ -wherein α is a settable parameter, the physical meaning is a color difference threshold;

step 2.2: selecting false same-color pairs, namely calculating all candidate colors, judging the threshold condition, and respectively selecting elements in a false same-color pair set according to a set color difference threshold alpha;

step 2.3: storing the pseudo-same-color pair, namely storing the selected pseudo-same-color set;

step 3: generating a test image, namely selecting a color blindness test pattern, filling a foreground and a background with a pseudo-same color pair to form a unit test pattern, and grouping the unit test pattern;

the step 3 comprises the following steps:

step 3.1: selecting a test pattern, namely selecting geometric shapes which are convenient to identify and common patterns in life as the test pattern, combining any false same-color pair generated in the process as a foreground color and a background color respectively, and filling the foreground and the background of the test pattern;

step 3.2: filling the test pattern, namely filling the test pattern by adopting the selected pseudo-same color pair;

step 3.3: generating a test unit, namely filling the test pattern to obtain a three-channel color image serving as the test unit;

step 3.4: the test images are grouped, i.e. a plurality of test units are selected randomly and combined into a group of test images.

2. The abnormal color vision test chart synthesizing method based on pseudo-same color search as defined in claim 1, wherein the abnormal color vision test chart synthesizing method is characterized by comprising the following steps:

in the step 1.1, dividing the RGB color space into grids of X multiplied by Y multiplied by Z, wherein the size is 32 multiplied by 32, and selecting a triplet candidate color vector on a node of the grid;

corresponding color vector Q in step 1.2 ^τ Is calculated as follows: firstly, converting the triplet candidate color vector Q into an LMS space to obtain a corresponding color vector in the LMS space, wherein the color vector is represented by the following formula:

wherein the method comprises the steps of

A transformation matrix for the RGB color space to the LMS color space;

and then using a color blindness simulation algorithm to obtain tau-type color blindness simulation vectors in the LMS color space:

wherein T is _sim ^τ The transformation matrix is used in the color blindness simulation algorithm;

the tau-color blind simulation vector of the LMS color space is then converted back into an RGB space vector, i.e.,

wherein Q is ^τ Corresponding colors, i.e., { Q, are simulated for achromatopsia of the triplet candidate color vector Q ^τ -a pair of pseudo-homochromes;

in step 1.3 { Q, Q is calculated according to the following formula ^τ Color difference value between }:

ΔE* _Lab (Q,Q ^τ )＝[(L _Q *-L _Qτ *) ² +(a _Q *-a _Qτ *) ² +(b _Q *-b _Qτ *) ² ]

wherein (L) _Q *,a _Q *,b _Q * ) Representing the result of conversion of candidate color Q into LAB space, (L) _Qτ *,a _Qτ *,b _Qτ * ) Representing the simulation color Q ^τ Three components after conversion to LAB space;

step 2.1 employs α=50;

step 2.2, randomly selecting in the selected pseudo-same color pair set, or directly picking up color elements in the set after visualization;

the storage mode in the step 2.3 can adopt a hash list or dictionary data structure, and a file format can use JSON or Pickle objects;

randomly generating numbers 0-9 as foreground contents and the rest blank area as background in the step 3.1 to form a test pattern;

in step 3.4, 5*5 unit test patterns are selected to form one test image.

3. The abnormal color vision test chart synthesizing method based on pseudo-same color search as defined in claim 2, wherein the abnormal color vision test chart synthesizing method is characterized in that:

in step 1.1, the RGB color space is divided into grids of X Y X Z, and the grids are divided into grid nodes (X ₀ ,y ₀ ,z ₀ ) As a center, the points reached by the random small step offsets (Δx, Δy, Δz) in each direction are used as triplet candidate color vectors.

4. The abnormal color vision test chart synthesizing method based on pseudo-same color search as defined in claim 2, wherein the abnormal color vision test chart synthesizing method is characterized in that:

in the step 3.2, the test pattern is directly filled with the pseudo-same color pair, the simulated color is used as the background filling color, and the original color is used as the foreground filling color.

5. The abnormal color vision test chart synthesizing method based on pseudo-same color search as defined in claim 2, wherein the abnormal color vision test chart synthesizing method is characterized in that:

and 3.2, filling the test pattern with the pseudo-same color pair split type, and drawing small circle geometric patterns which are not overlapped with each other by adopting a K-D tree data structure to form a split type test pattern unit.

6. The abnormal color vision test chart synthesizing method based on pseudo-same color search as defined in claim 2, wherein the abnormal color vision test chart synthesizing method is characterized in that:

in step 3.2, the test pattern is filled with the approximate color generalization of the pseudo-same color pair, and similar colors are randomly picked up according to Gaussian distribution in a sphere space with the pseudo-same color as a sphere center and the radius rho in an LAB color space to fill small geometric patterns generated each time, wherein rho is more than 0 and less than 32.

7. The abnormal color vision test chart synthesizing method based on pseudo-same color search as defined in claim 3, wherein the abnormal color vision test chart synthesizing method is characterized in that: step 1.1 taking a die of small step size

Wherein Δx, Δy, Δz are integers;

the test pattern is filled in a fine density checkered pattern in step 3.2.

8. An abnormal color vision test chart application device is characterized in that: the image generated by the false same color search based abnormal color vision test chart synthesis method according to any one of claims 1 to 7 can be used for a special color vision detection computer device, embedded software and hardware, paper printing device or internet application software system.

9. The abnormal color vision test chart application apparatus according to claim 8, wherein: the color vision detection computer device comprises various types of computers, mobile terminals, virtual Reality (VR) equipment such as Oculus, google glasses, iPad tablet computers, special displays and medical eye vision tester equipment, wherein the method for synthesizing the color vision test patterns is realized by using the off-line or on-line application of the development of the corresponding medical eye vision tester equipment in a software system, and the color vision test patterns are presented on a display screen by taking unit test patterns and grouping test patterns as use cases; the embedded software and hardware and paper printing device realizes a color vision test chart synthesis method by a corresponding programming language through a singlechip, various embedded systems of an FPGA and a development board, or synthesizes color vision test charts on other hardware platforms in advance, and then stores the color vision test charts in embedded equipment for display or printing; the Internet application software system develops and realizes the color vision test synthesis method in various programming languages such as Javascript and hypertext markup language, or presents a color vision test chart at a browser or APP end in a micro-service and off-line application mode and combines with an interactive design to implement color vision test.

Images