CN112165613B

CN112165613B - Method for testing stereo sharpness of dynamic random point stereogram based on horizontal motion

Info

Publication number: CN112165613B
Application number: CN202010822106.5A
Authority: CN
Inventors: 沈丽丽; 刘志建; 侯春萍
Original assignee: Tianjin University
Current assignee: Tianjin University
Priority date: 2020-08-15
Filing date: 2020-08-15
Publication date: 2021-09-07
Anticipated expiration: 2040-08-15
Also published as: CN112165613A

Abstract

The invention relates to a method for testing the stereo sharpness of a dynamic random point stereogram based on horizontal motion, which comprises the following steps: material manufacturing and parameter selection; and (3) material presentation: after selecting proper point size, point density and motion speed parameters, making video form dynamic random point diagrams with different parallaxes, different presentation modes including cross parallaxes and non-cross parallaxes and horizontal movement with C-shaped stereoscopic effect in different opening directions; the material is divided into three parts in the test, and according to a gradient descending mode, the first part is as follows: a high parallax portion; a second part: a medium parallax component; and a third part: a low-disparity portion, the three portions being presented in order, each portion being presented randomly internally; requiring the subject to make a decision during testing; data processing: the judgment made by the tested object is collected during the test period, and can be extracted after the test is finished, and the summarized data obtains the judgment accuracy of the tested object under each parallax; and performing smooth spline interpolation fitting on the data, wherein each tested data can be fitted with a curve of the accuracy and the parallax, and the parallax corresponding to a point with the accuracy of 75% on the curve is the tested stereoscopic acuity threshold value.

Description

Method for testing stereo sharpness of dynamic random point stereogram based on horizontal motion

The technical field is as follows:

the present invention relates to the field of ophthalmic stereoscopic acuity testing.

Background art:

the stereoscopic vision is the ability of distinguishing and perceiving the three-dimensional space positions of the distance, depth and height of surrounding objects by visual organs, and is an independent high-grade binocular vision function established on the basis of binocular simultaneous vision and fusion function. The better the stereo vision, the stronger the ability to judge distance finely, and the easier it is to detect subtle differences in distance. For the occupations of automobile driving, precise instrument operation, aviation and navigation and the like, good stereoscopic vision is an important condition for finishing work excellently. The unit of stereo sharpness (stereoacuity) used for evaluating stereo vision is commonly referred to as arc seconds, or angular seconds (arcsec, "), and the size of the unit is determined by measuring the minimum parallax (disparity) perceived by two eyes, and the smaller the threshold value of the parallax is, the better the stereo sharpness is. The common measurement methods are stereographic inspection based on polarized glasses and random point stereoscopy based on red green (or blue) filters.

Some common methods for testing the stereo sharpness have the problems of poor measurement accuracy, easy occurrence of errors, limited measurement range and the like. For example, the Howard-Dolman Test (Howard-Dolman Test) is an ancient method for testing the sharpness of stereoscopic vision, and the method is composed of two Test rods, wherein one Test rod is fixed in position, and the other Test rod can move freely. The test was run using a rope to manipulate the travel bars until both were felt to be at the same horizontal distance. The sharpness of vision is calculated using a formula based on the distance between the two rods in the vertical direction. This method is complicated to operate and is easily interfered by various external factors, so that it is inconvenient for clinical use. The Frisby stereoscopic vision Test (Frisby Stereo Test) is performed by three sets of glass plates having different thicknesses, and requires moving and turning the glass plates during the Test, and thus is not convenient in clinical tests. The three-dimensional Fly Test (Stereo Fly Test) adopts a polarization technology. During testing, polarized light eyes are needed to observe the matched polarized pictures, the testing time difference is respectively 100 ', 200 ' and 300 ', the testing step length is large, and the precision is limited. The TNO Test (TNO Test) uses a red-green filter technique, with a pattern consisting of red and green. When a tested object needs to wear red and green glasses, on the basis of the color complementation principle, the red lens on the left side can only see red patterns, and the green lens on the right side can only see green patterns. When a test is carried out, a three-dimensional shape represented by a red-green random dot diagram corresponding to a shape visible to a single eye needs to be pointed out. The TNO test step length is larger, a more accurate value cannot be obtained, and the red and green glasses have poorer light transmission, so the observation of the three-dimensional effect is greatly discounted.

Most of the existing three-dimensional testing methods often have the problems of large steps, incapability of accurately testing the visual acuity, inconvenience in operation, unsuitability for clinical tests and the like. Furthermore, most of the stereoscopic vision tests are static tests, and dynamic observation is often accompanied in daily life and work of people. It is still necessary to develop a dynamic acuity testing method which has a small step size, relatively accurate measurement results, can measure dynamic acuity and is easy to introduce into clinical tests.

The invention content is as follows:

the invention provides a method for testing the stereoscopic acuity under the dynamic condition aiming at the stereoscopic acuity test, which has the technical scheme that:

a method for testing the stereo sharpness of a dynamic random point stereogram based on horizontal motion comprises the following steps:

the first step is as follows: material making and parameter selection:

1) and (3) parallax calculation: calculating the relation between parallax and viewing distance according to a binocular imaging principle; selecting a proper observation distance;

2) material preparation: generating a random point stereogram pair with controllable point size, point density and random point graph size; making a graph pair suitable for the 3D display according to the resolution and the point distance parameters of the 3D display; making a dynamic random point stereogram, and displaying the dynamic random point stereogram in a video form or an kinegram form;

3) parameter selection: selecting the size and the point density range of the random point stereogram, testing to select the parameters of the random point stereogram with the best effect, making random point stereograms with different speeds, and testing to select a proper movement speed;

the second step is that: presentation of material

After selecting proper point size, point density and motion speed parameters, making video form dynamic random point diagrams with different parallaxes, different presentation modes including cross parallaxes and non-cross parallaxes and horizontal movement with C-shaped stereoscopic effect in different opening directions; the material is divided into three parts in the test, and according to a gradient descending mode, the first part is as follows: a high parallax portion; a second part: a medium parallax component; and a third part: a low-disparity portion, the three portions being presented in order, each portion being presented randomly internally; requiring the subject to make a decision during testing;

the third step: data processing

1) The judgment made by the tested object is collected during the test period, and can be extracted after the test is finished, and the summarized data obtains the judgment accuracy of the tested object under each parallax;

2) and performing smooth spline interpolation fitting on the data, wherein each tested data can be fitted with a curve of the accuracy and the parallax, and the parallax corresponding to a point with the accuracy of 75% on the curve is the tested stereoscopic acuity threshold value.

Description of the drawings:

through the attached drawings (tables), the implementation steps and the advantages of the invention can be more intuitive, and meanwhile, the reader can more easily understand the flow and the operation of the invention.

FIG. 1 is a cross-disparity diagram

FIG. 2 is a non-cross parallax schematic diagram

Different presentation of the random dot diagram in fig. 3 a random dot diagram pair (2D display) b random dot diagram presents effect C on the 3D display, the human-observable stereoscopic effect (effect seen by human eyes), the cross-parallax presentation effect is that C composed of one random dot floats on the background random dot diagram

FIG. 4 shows data obtained by fitting different persons, wherein a and b are curves obtained by fitting data of two persons to be tested

TABLE 1 RDS parameters tested for selection

TABLE 2 DRDS parameters tested for selection

TABLE 3 test results of the sharpness of stereoscopic vision

The specific implementation mode is as follows:

in order to make the technical solution of the present invention more clear and easy to implement, so as to further highlight the advantages and objects of the present invention, the embodiments of the present invention will be further described and explained in detail with reference to the accompanying drawings.

The invention provides a method for testing the stereo sharpness, which is a method for testing the stereo sharpness under the dynamic condition, namely a method for testing the stereo sharpness based on a dynamic random point diagram. The method generates a Dynamic random-dot stereogram (DRDS) with a C-type stereo effect through programming, wherein the C-type stereo has four openings respectively: the parallax of the DRDS and the observation distance of the DRDS are calculated rigorously, and the DRDS is displayed on a 3D display and needs to be observed by polarized 3D glasses. The method aims at the parameters of DRDS such as: the factors such as point density, point size, movement speed and the like are researched, and appropriate parameters are selected to ensure the accuracy of the experimental result. The presentation of the stimulation is carried out by adopting a gradient descending method, the test is divided into three parts according to gradient change, and each part presents dynamic random point stereograms with a plurality of parallaxes. In each test, the tested three-dimensional effect needs to be judged, the judgment results can be collected uniformly after the test is finished, the judgment results are analyzed through programming, the judgment accuracy of the tested under each parallax condition is summarized, smooth spline interpolation is used for fitting data, a curve of the tested accuracy judgment rate and the parallax is fitted, the parallax corresponding to the position with the accuracy of 75% is selected as the test result, and the test result is the tested dynamic three-dimensional visual acuity threshold value.

101: calculation of disparity

The brain extracts depth information in stereoscopic vision using binocular parallax, thereby forming depth perception. The perceived objects may appear on the stereoscopic display screen, or in front of or behind the display screen, and thus the resulting binocular Disparity is also referred to as Zero Disparity (Zero Disparity), cross Disparity (cross Disparity), and non-cross Disparity (cross Disparity), respectively. As shown in fig. 1 and 2:

s in formula (1) is the observation distance, and the value of the relative displacement l to the point a and the point B, which can be calculated by formula (1), is in millimeters. For ease of programming in matlab, equation (1) is transformed as shown below:

in the formula (2), V is a parallax value (unit arc second), Deltal is a pixel value of the movement of a corresponding point in a random point diagram pair, and d₁The point distance of the 3D display is shown, and S is the observation distance to be tested. The 3D display used in the test had a resolution of 1920 x 1080px and a dot spacing of 0.248mm/px, with S being 102.3cm for ease of programming according to equation (2).

102: dynamic random point map parameter setting

Research data show that the parameters of the random dot diagram influence the stereoscopic effect of the random dot diagram. And aiming at the dynamic random point diagram, selecting parameters such as point size, point density, movement speed and the like for research so as to select proper parameters and ensure the observation effect of the dynamic random point stereogram.

1) Dot size, dot density

Many documents show that the dot size and the dot density of the random dot pattern have a great influence on the effect of the random dot pattern. Therefore, it is important to select appropriate parameters. The selection of the dot size and the density is divided into two modules, the size and density selection range is reduced through static random dot diagram test, and parameters are determined through dynamic random dot diagram test.

According to the prior art, a static random dot diagram (RDS) dot density range is selected as follows: 0.2, 0.25, 0.3, 0.35, 0.4, 0.45, 0.5; dot size range: 2.3, 4, 5, 6 (unit px). The static random dot plot disparity values used for the test were: 80",140". The images were divided into four groups, a first group size of 4px with parallax 80 ", a second group size of 4px with parallax 140", a third group density of 0.3 with parallax 80 ", and a fourth group density of 0.3 with parallax 140". The first two groups select the density of the graph with the best three-dimensional effect under the condition of determining parallax and size; the latter two groups select the parallax of the image with the best three-dimensional effect under the condition of determining the density and the size. The test results are shown in table 1. The size range is limited to 2, 3, 4, 5 (unit px), and the density range is limited to 0.3, 0.35, 0.4.

The size and density were further determined using dynamic random point testing with the parallax set to 80 ". The experiment is divided into four groups, wherein the first group of dynamic random dot diagram has a parallax of 80 ″, and the density of 0.3 is 2, 3, 4 and 5 (unit px); a first set of dynamic random dot patterns with a disparity of 80 "and a density of 0.35 of 2, 3, 4, 5 (in px); a first set of dynamic random dot patterns with a disparity of 80 "density 0.4 of 2, 3, 4, 5 (in px); and selecting one random dot diagram with the best effect from the three groups to form a fourth group, and selecting a dynamic random dot diagram with the best three-dimensional effect from the fourth group. The final spot size selected over multiple tests was 3 × 3px with a spot density of 0.3. The test results are shown in table 2.

2) Moving speed of three-dimensional figure

The velocity also has an effect on the stereoscopic effect of the random dot pattern. The display presents a random dot diagram area of 600 × 600px, the outer ring radius of the stereograph C is 200px, the inner ring radius is 150px, the area of C occupies about 30% of the area of the random dot diagram area, the viewing distance is 102.3cm, and seven grades of speed are selected for determining the speed parameter: 0.279 DEG/s, 0.558 DEG/s, 1.116 DEG/s, 2.232 DEG/s, 4.464 DEG/s, 8.928 DEG/s, 17.856 DEG/s. The parallax ranges are 20 ", 30", 40 ". The tested object can observe the DRDS with different horizontal movement speeds under each parallax in sequence, and the best DRDS movement speed suitable for testing is selected according to the feedback of a plurality of tested objects.

The feedback to be tested was at speed levels five, six, seven: the observation effects of 4.464 degrees/s, 8.928 degrees/s and 17.856 degrees/s are poor, the movement speed of the stereo graph is high, the stereo effect is not easy to observe, and the observed parallax can be clearly increased gradually along with the increase of the speed; speed class three, four: 1.116 degrees/s and 2.232 degrees/s, the observation effect of the speed at the level is better than that of other speed grades under the same parallax level; speed class one, two: 0.279/s, 0.558/s, which is better than the high speed (five, six, seven) but worse than the medium speed (three, four). According to the feedback of the test, the DRDS speed for the test is between 1.116 DEG/s and 2.232 DEG/s and is about 1.5 DEG/s.

In summary, the point density of the dynamic random point stereogram is 0.3, the point size is 3 × 3px, the observation distance is 102.3cm, and the horizontal movement speed is 1.5 °/s. The random point perspective view is shown in fig. 3, and the random point views exist in pairs, and in the 2D case, are shown in fig. 3 (a); presented via a 3D display as shown in figure (b); after wearing the 3D polarized glasses, the observation effect is as shown in fig. 3(C) (a random point perspective view in the case of cross parallax), and there is a stereoscopic "C" character effect, floating on the background random point.

103: stimulus presentation

When the material is presented, the dynamic random point diagrams with different parallax sizes are presented in a gradient descending mode. The 19 levels of disparity are grouped by the size of disparity.

The material is presented and divided into three parts:

a first part: 45 ", 50", 60 ", 70", 80 ", 90", 100 ", 120", 200 ", 300", 400 ";

a second part: 25 ", 30", 35 ", 40", 45 ";

and a third part: 5 ', 10', 15 ', 20'

The three parts are sequentially presented in sequence, and in each part, dynamic random point stereograms with different parallaxes and different opening directions are presented randomly. The respective parallax levels of the first part will be presented 8 times with different opening directions, intersecting non-intersecting differences, of which three parallaxes of 45 ", 50", 60 "are presented 16 times. The dynamic random dot stereograms corresponding to the second partial parallax are each presented 24 times, and the first and second portions are presented 16 times with an intersection 45 ". The dynamic random point stereograms corresponding to the third partial parallax are presented 16 times each.

Each trial time shows a picture of focusing attention for 0.5s, and then shows a dynamic random point stereo image of horizontal movement for 1s, and the trial needs to make judgment according to the opening direction key of the stereo image. If and only if a decision is made by the trial, then the next trial will be skipped. The decision made at each trial is recorded.

104: data processing

The judgment results of the opening directions of the materials to be tested are summarized, the data are quickly processed by software programming, the selection made by the test can be correctly and quickly judged, and the judgment accuracy of the test corresponding to each parallax is obtained. The judgment accuracy of each parallax is processed and fitted by a smooth spline (realized by software programming) mode to form a curve of the judgment accuracy with respect to the parallax, and a parallax value corresponding to a point with the accuracy of 75% is found, namely the tested stereoscopic vision sharpness threshold. As shown in fig. 4, is a curve fitted to the results of two tests tested.

TABLE 1 RDS parameters tested for selection

TABLE 2 DRDS parameters tested for selection

TABLE 3 test results of the sharpness of stereoscopic vision

After a plurality of tests, the test results are shown in table 3; the stereo visual acuity testing method based on the dynamic random point diagram can accurately test the tested visual acuity; and the test method is simple and easy to operate. The dynamic random point stereogram of the horizontal movement of different openings with different parallaxes can be watched in a few minutes before the display is tried on, and the judgment is made, so that the accurate stereo acuity can be obtained, and the method has the advantages of better practicability, easiness in use and suitability for popularization.

Claims

1. A method for testing the stereo sharpness of a dynamic random point stereogram based on horizontal motion comprises the following steps:

the first step is as follows: material making and parameter selection:

2) material preparation: generating a random point stereogram with controllable point size, point density and random point graph size; manufacturing a dynamic random point stereogram suitable for the 3D display according to the resolution and the point distance parameters of the 3D display, and displaying the dynamic random point stereogram in a video form or an kinegram form;

3) parameter selection: selecting the point size and point density range of the random point stereogram, testing to select the parameters of the random point stereogram with the best effect, making random point stereograms with different speeds, and testing to select a proper movement speed;

the second step is that: presentation of material

After selecting proper point size, point density and motion speed parameters, making video form dynamic random point stereograms with different parallaxes, different presentation modes including cross parallaxes and non-cross parallaxes and different opening directions and having C-shaped stereo effect horizontal movement; the material is divided into three parts in the test, and according to a gradient descending mode, the first part is as follows: a high parallax portion; a second part: a medium parallax component; and a third part: a low-disparity portion, the three portions being presented in order, each portion being presented randomly internally; requiring a testee to judge the opening direction during the test period, and recording judgment data;

the third step: data processing

1) The judgment correctness and mistake made by the testee during the test period can be collected, the judgment correctness and mistake can be extracted after the test is finished, and the judgment correctness of the testee under each parallax is obtained by summarizing data;

2) and performing smooth spline interpolation fitting on the data, wherein a curve with the accuracy changing along with the parallax can be fitted to the data of each tested person, and the parallax corresponding to the point with the accuracy of 75% on the curve is the stereoscopic vision sharpness threshold of the tested person.