Image-background separation spectral depth perception evaluation training method

Abstract

The present application provides a kind of image-background separation spectral depth perceptual evaluation training method, it is related to the technical field of depth perceptual evaluation, in the background canvas cooperation, the foreground image model corresponds to different spectral depth concave-convex relationship, the foreground image model observed and fed back by observer through observation equipment, the concave-convex relationship observed is namely the spectral depth concave-convex relationship corresponding to the foreground image model that reflects the depth perceptual perception of itself, the concave-convex relationship of the foreground image model observed and fed back by observer through observation equipment is compared with the spectral depth concave-convex relationship corresponding to the foreground image model set in advance, the contour integration function condition of the spectral depth perceptual of observer is obtained, the scheme helps observer to find the abnormal situation of spectral depth perceptual of itself in time, then through training, the depth perceptual contour integration function of observer is assisted to improve, fill the blank of the current research in this aspect, and it has strong guiding significance to clinical depth perceptual function.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of depth perception evaluation, and more particularly to a spectral depth perception evaluation training method of image-background separation. BACKGROUND

[0002] Contour integration is the ability of an observer to segment a visual image and perceive the contour boundary, which mainly integrates local features in the field of view, such as direction, depth, color, etc. In a two-dimensional processing mechanism, contour integration requires the integration of local elements on a global contour image, which may be directional, static orientation, color, texture, disparity, etc. Traditional research believes that contour integration is a two-dimensional process, that is, the process of connecting directional segments along the direction axis. However, contour lines also exist in the three-dimensional world. If the contour and the noise background are at different depths, the ability of an observer with normal stereoscopic vision to separate the collinear contour from the noise background will be stronger. Therefore, in people's life, contour integration and human depth perception have a certain inevitable relationship.

[0003] Image-background separation is the process of dividing a visual image into an image and a background by the visual system, which is the first step of visual perception. In the three-dimensional world, the image-background separation process is the process of observer's depth perception cooperation, in which the spectral and color signals in the visual image are an important embodiment in the image-background separation. In the spectral color system, different colors have different wavelengths of light. When the lens does not focus all wavelengths of light at the same point, chromatic aberration occurs. Different wavelengths are refracted at different angles on the lens of the human eye. This causes the focal point of each color to be at different positions on the retina, forming different depth perceptions.

[0004] Images of the same color system have similar depth perception and are integrated into an image contour by the visual system. For example Figure 1 The red-green color blindness detection chart shown in the figure, the color of the background in the chart basically belongs to the red spectrum, and is imaged behind the retina of the observer. The color of "7" basically belongs to the green spectrum, and the imaging position is in front of the red spectrum. Therefore, a normal observer observing the red-green color blindness detection chart forms a whole contour of "7" protruding from the red spectrum background. Figure 1 The red-green color blindness detection chart shown in the figure, the color of the background in the chart basically belongs to the red spectrum, and is imaged behind the retina of the observer. The color of "7" basically belongs to the green spectrum, and the imaging position is in front of the red spectrum. Therefore, a normal observer observing the red-green color blindness detection chart forms a whole contour of "7" protruding from the red spectrum background. Figure 1The red-green color blindness detection diagram shown, in the conventional visual color blindness inspection, the color blindness crowd cannot see the color, therefore, the outline of the image cannot be distinguished, in fact, it is a contour integration function abnormality in the spectral depth perception, and this aspect is still relatively blank in the current research, therefore, the spectral depth perception evaluation training of the image-background separation of people has guiding significance for the improvement of the function of the contour integration of the depth perception in the later stage. SUMMARY

[0005] In order to solve the problem of how to carry out the spectral depth perception evaluation training of the image-background separation of people, the present application provides a spectral depth perception evaluation training method of image-background separation, which helps the observer to discover the abnormal condition of the spectral depth perception in time, and assists in improving the contour integration function of the depth perception of the observer, fills the blank of the current research in this aspect, and has strong guiding significance for clinical vision.

[0006] In order to achieve the above technical effects, the technical scheme of the present application is as follows:

[0007] A spectral depth perception evaluation training method of image-background separation, the evaluation training method comprises:

[0008] S1. Setting the upper limit of the picture content and the observation evaluation times as N, the picture content comprising a background canvas and a foreground image model, the foreground image model corresponding to different spectral depth concave-convex relationships under the cooperation of the background canvas;

[0009] S2. Displaying the picture content through a display device, and the observer observing the picture content displayed in the display device through an observation equipment and feeding back the concave-convex relationship of the foreground image model in the observed picture content;

[0010] S3. Judging whether the concave-convex relationship of the foreground image model fed back by the observer is consistent with the spectral depth concave-convex relationship corresponding to the foreground image model set in advance, if yes, executing step S4; otherwise, the spectral depth perception of the observer is abnormal, executing step S5;

[0011] S4. Judging whether the observation evaluation times reach N, if yes, the spectral depth perception of the observer is normal; otherwise, returning to step S2;

[0012] S5. Setting the total training period and the display state of the picture content in the display device in each training period, and carrying out the spectral depth perception training on the observer in each training period.

[0013] In the technical solution, the cooperation of image-background separation and observer depth perception is considered, the picture content is set, and the picture content includes a background canvas and a foreground image model. Under the cooperation of the background canvas, the foreground image model corresponds to different spectral depth concave-convex relationships. The foreground image model observed by the observer through the observation equipment and fed back is the foreground image model corresponding to the spectral depth concave-convex relationship perceived by the observer, which is the embodiment of the observer's depth perception. The concave-convex relationship of the foreground image model observed by the observer through the observation equipment and fed back is compared with the spectral depth concave-convex relationship corresponding to the foreground image model set in advance. If the concave-convex relationship of the foreground image model observed and fed back is consistent with the spectral depth concave-convex relationship corresponding to the foreground image model set in advance under N times of evaluation, the observer's spectral depth perception is normal, otherwise, the observer's spectral depth perception is abnormal. The scheme helps the observer to find the abnormal condition of the observer's spectral depth perception in time, and then the observer's depth perception profile integration function is improved through training.

[0014] Preferably, the background canvas in the picture content is a solid color background or a bar grid or a checkered background; the width of the background canvas is 2-3 times the width of the foreground image model, that is, the range of the background canvas is larger than the range of the foreground image model.

[0015] Preferably, in step S2, when the picture content is displayed by the display device, the color of the background canvas in the picture content is random, and the foreground image model is also random, but when the color of the background canvas changes from A color to B color, the foreground image model cooperating with the A color background canvas is the same as the foreground image model cooperating with the B color background canvas.

[0016] In the spectral color system, different colors have different wavelengths, so in the process of changing the color of the background canvas, the variable control in the evaluation process is considered to ensure the reliability of the evaluation result, and the foreground image model cooperating with the A color background canvas is the same as the foreground image model cooperating with the B color background canvas.

[0017] Preferably, in the picture content, each component of the foreground image model is static and does not block each other, so as to avoid interference of the feedback concave-convex relationship.

[0018] Preferably, the picture content is that a foreground image model is arranged on a white background canvas, and the foreground image model includes a green circle and a red animal in the green circle; or the picture content is that a foreground image model is arranged on a black background canvas, and the foreground image model includes a green circle and a red animal in the green circle; under the cooperation of the background canvas of different colors, the red animal in the green circle in the foreground image model has a depth concave-convex relationship with the green circle.

[0019] Preferably, in step S2, the observation device is cross 3D spectral separation glasses, if the background canvas is white, the green circle and the foreground image model of the red animal located in the green circle are set on the background canvas, the pre-set spectral depth concave-convex relationship corresponding to the red animal in the foreground image model is "concave", if the observer observes the picture content displayed in the display device through the observation device and feeds back that the position of the red animal is "concave", the concave-convex relationship of the foreground image model fed back by the observer is consistent with the pre-set spectral depth concave-convex relationship corresponding to the foreground image model; otherwise, it is inconsistent.

[0020] If the background canvas is a black background, the green circle and the foreground image model of the red animal located in the green circle are set on the background canvas, the pre-set spectral depth concave-convex relationship corresponding to the red animal in the foreground image model is "convex", if the observer observes the picture content displayed in the display device through the observation device and feeds back that the position of the red animal is "convex", the concave-convex relationship of the foreground image model fed back by the observer is consistent with the pre-set spectral depth concave-convex relationship corresponding to the foreground image model; otherwise, it is inconsistent.

[0021] In step S4, if the observer's spectral depth perception is normal, in N observation evaluations, the concave-convex relationship of the foreground image model fed back by the observer is consistent with the pre-set spectral depth concave-convex relationship corresponding to the foreground image model each time.

[0022] Preferably, in step S2, the observation device is non-cross 3D spectral separation glasses, if the background canvas is white, the green circle and the foreground image model of the red animal located in the green circle are set on the background canvas, the pre-set spectral depth concave-convex relationship corresponding to the red animal in the foreground image model is "convex", if the observer observes the picture content displayed in the display device through the observation device and feeds back that the position of the red animal is "convex", the concave-convex relationship of the foreground image model fed back by the observer is consistent with the pre-set spectral depth concave-convex relationship corresponding to the foreground image model; otherwise, it is inconsistent.

[0023] If the background canvas is a black background, the green circle and the foreground image model of the red animal located in the green circle are set on the background canvas, the pre-set spectral depth concave-convex relationship corresponding to the red animal in the foreground image model is "concave", if the observer observes the picture content displayed in the display device through the observation device and feeds back that the position of the red animal is "concave", the concave-convex relationship of the foreground image model fed back by the observer is consistent with the pre-set spectral depth concave-convex relationship corresponding to the foreground image model; otherwise, it is inconsistent.

[0024] In step S4, if the observer's spectral depth perception is normal, in N observation evaluations, the concave-convex relationship of the foreground image model fed back by the observer is consistent with the pre-set spectral depth concave-convex relationship corresponding to the foreground image model each time.

[0025] The above two cases of intersection / non-intersection evaluate the spectral depth perception from different angles, and guarantee the reliability of the evaluation results.

[0026] Preferably, in the step S5, when setting the display state of the picture content in the display device in each training period, the setting includes setting: the color of the background canvas, the shape composition of the foreground image model under the color, and the continuous presentation time of the canvas content formed by the combination of the color background canvas and the foreground image model under the color in each training period.

[0027] When training the spectral depth perception of the observer, the following training steps are included:

[0028] S101. In each training period, the color of the background canvas is set to be unchanged or changed, the picture content is displayed in the display device according to the set continuous presentation time, and the observer feeds back the concave-convex relationship of the foreground image model through the observation equipment.

[0029] S102. Determine whether the concave-convex relationship of the foreground image model fed back by the observer is consistent with the spectral depth concave-convex relationship corresponding to the foreground image model set in advance, if yes, increase the shape transformation frequency of the foreground image model until the total training period is completed, otherwise, execute step S103.

[0030] S103. Reduce the shape transformation frequency of the foreground image model and return to step S102.

[0031] The present application also proposes an image-background separation spectral depth perception evaluation training system, the evaluation training system includes:

[0032] A display device is used to display picture content, and the picture content includes a background canvas and a foreground image model, and the foreground image model corresponds to different spectral depth concave-convex relationships under the cooperation of the background canvas.

[0033] An observation equipment is used when the observer observes the picture content displayed in the display device, and the observer determines the concave-convex relationship of the foreground image model in the viewed picture content through the observation equipment.

[0034] A judgment unit is used to determine whether the concave-convex relationship of the foreground image model fed back by the observer is consistent with the spectral depth concave-convex relationship corresponding to the foreground image model set in advance.

[0035] Compared with the prior art, the technical scheme of the present application has the following beneficial effects:

[0036] The present application provides a kind of image-background separation spectral depth perception evaluation training method and system, in the background canvas cooperation, the foreground image model corresponds to different spectral depth concave-convex relationship, observer observes and feedbacks the foreground image model by observation equipment, the concave-convex relationship observed is namely the spectral depth concave-convex relationship corresponding to the foreground image model of the depth perception perception of self, the concave-convex relationship of the foreground image model observed and feedbacked by observation equipment is compared with the spectral depth concave-convex relationship corresponding to the foreground image model set in advance, the profile integration function condition of the spectral depth perception of observer is obtained, the scheme helps observer to find the abnormal situation of self spectral depth perception in time, then through training, it is assisted to improve the depth perception profile integration function of observer, fills the blank of current research in this aspect, has strong guiding significance to clinical vision. BRIEF DESCRIPTION OF DRAWINGS

[0037] Figure 1 The red-green color blindness detection diagram mentioned in the background art of the present application is shown;

[0038] Figure 2 The flowchart of the image-background separation spectral depth perception evaluation training method proposed in embodiment 1 of the present application is shown;

[0039] Figure 3 The schematic diagram of the picture content of the first background canvas color proposed in embodiment 1 of the present application is shown;

[0040] Figure 4 The schematic diagram of the picture content of the second background canvas color proposed in embodiment 1 of the present application is shown;

[0041] Figure 5 The schematic diagram of the picture content of the third background canvas color proposed in embodiment 1 of the present application is shown;

[0042] Figure 6 The flowchart of the spectral depth perception training for observer proposed in embodiment 1 of the present application is shown;

[0043] Figure 7 The structural schematic diagram of the image-background separation spectral depth perception evaluation training system proposed in embodiment 3 of the present application is shown. DETAILED DESCRIPTION

[0044] The drawings are only used for illustrative description, and cannot be understood as limitation to the present patent;

[0045] In order to better illustrate the present embodiment, some parts of the drawings may be omitted, enlarged or reduced, and do not represent the actual size;

[0046] For those skilled in the art, it is understandable that some well-known content in the drawings may be omitted.

[0047] The technical solutions of the present application are further described below in combination with the drawings and embodiments.

[0048] The position relationship described in the drawings is only for exemplary illustration and cannot be understood as a limitation of the present patent;

[0049] Embodiment 1

[0050] As Figure 2 shown, the present embodiment proposes an image-background separation spectral depth perception evaluation training method, see Figure 2 , which comprises:

[0051] S1. Set the upper limit of the picture content and the observation evaluation times as N, the picture content comprises a background canvas and a foreground image model, and the foreground image model corresponds to different spectral depth concave-convex relationships under the cooperation of the background canvas;

[0052] S2. Display the picture content through a display device, and an observer observes the picture content displayed in the display device through an observation equipment and feeds back the concave-convex relationship of the foreground image model in the observed picture content;

[0053] S3. Determine whether the concave-convex relationship of the foreground image model fed back by the observer is consistent with the pre-set spectral depth concave-convex relationship corresponding to the foreground image model, if yes, execute step S4; otherwise, the observer's spectral depth perception is abnormal, execute step S5;

[0054] S4. Determine whether the observation evaluation times reaches N, if yes, the observer's spectral depth perception is normal; otherwise, return to step S2;

[0055] S5. Set the total training period and the display state of the picture content in the display device in each training period, and perform spectral depth perception training on the observer in each training period.

[0056] Overall, considering the cooperation of image-background separation and observer depth perception, the picture content is set, which includes a background canvas and a foreground image model. Under the cooperation of the background canvas, the foreground image model corresponds to different spectral depth concave-convex relationships. The concave-convex relationship observed by the observer through the observation equipment and the feedback is the spectral depth concave-convex relationship corresponding to the foreground image model perceived by the observer's depth perception, which is the embodiment of the observer's depth perception. The concave-convex relationship of the foreground image model observed by the observer through the observation equipment and the feedback is compared with the spectral depth concave-convex relationship corresponding to the foreground image model set in advance. Under N times of evaluation, if the concave-convex relationship of the foreground image model observed and fed back is consistent with the spectral depth concave-convex relationship corresponding to the foreground image model set in advance, the observer's spectral depth perception is normal, otherwise, the observer's spectral depth perception is abnormal. In this embodiment, N is 3. When the concave-convex relationship is consistent for 3 times, it means that the observer's spectral depth perception is normal, otherwise, the observer's spectral depth perception is abnormal. This scheme helps the observer to discover the abnormal situation of his own spectral depth perception in time, and then through training, it helps to improve the observer's depth perception contour integration function.

[0057] In this embodiment, the background canvas in the picture content is a pure color background or a bar grid or a square grid background. The width of the background canvas is 2-3 times the width of the foreground image model, that is, the range of the background canvas is larger than the range of the foreground image model. When the picture content is displayed by the display device in step S2, the color of the background canvas in the picture content is random, and the foreground image model is also random. However, when the color of the background canvas changes from A color to B color, the foreground image model cooperating with the A color background canvas is the same as the foreground image model cooperating with the B color background canvas. Because in the spectral color system, the wavelengths of different colors are different, when the color of the background canvas changes, the variable control in the evaluation process is considered to ensure the reliability of the evaluation result, and the foreground image model cooperating with the A color background canvas is the same as the foreground image model cooperating with the B color background canvas.

[0058] In the picture content, each component of the foreground image model is static and does not block each other, so as to avoid interference of the feedback concave-convex relationship. Figures 3-5 The following are schematic diagrams of different picture contents, wherein, Figure 3 The picture content shown in the first background canvas color is black, and the foreground image model is different "circles" and animal images in the "circles", such as elephants, wolves and cats. Figure 4 The picture content shown in the second background canvas color is white, and the foreground image model is different "circles" and animal images in the "circles", such as elephants, wolves and cats. Figure 5Picture content representing a third background canvas color, the background canvas color being gray, and the foreground model being a wind rose, the three picture contents, the background canvas color of which is a pure color. Among them, the white background background canvas is provided with a foreground image model, and the foreground image model includes a green circle and a red animal in the green circle Figure 3 , or the black background background canvas is provided with a foreground image model, and the foreground image model includes a green circle and a red animal in the green circle Figure 4 , under the cooperation of background canvases of different colors, the red animal in the green circle in the foreground image model has a depth concave-convex relationship with the green circle.

[0059] 3D spectrum separation glasses are a kind of glasses that present depth perception based on colorful diffraction technology and special holographic light separation lenses. 3D spectrum separation glasses can effectively enhance the light separation aberration effect, and present different depth perception in color at different positions of the retina, such as more convex red series and more concave blue series. In the traditional case, 3D spectrum separation glasses are mostly used for entertainment. In step S2 of the embodiment, the observation device is cross 3D spectrum separation glasses, as shown in Figure 3 , if the background canvas is white, the background canvas is provided with a foreground image model of a green circle and a red animal (elephant) in the green circle, and the pre-set spectral depth concave-convex relationship of the red animal (elephant) in the foreground image model is "concave", if the observer observes the picture content displayed in the display device through the observation device and feeds back that the position of the red animal (elephant) is "concave", the concave-convex relationship of the foreground image model fed back by the observer is consistent with the pre-set spectral depth concave-convex relationship of the foreground image model; otherwise, it is inconsistent;

[0060] If the background canvas is a black background, as shown in Figure 4 , the background canvas is provided with a foreground image model of a green circle and a red animal (elephant) in the green circle, and the pre-set spectral depth concave-convex relationship of the red animal (elephant) in the foreground image model is "convex", if the observer observes the picture content displayed in the display device through the observation device and feeds back that the position of the red animal (elephant) is "convex", the concave-convex relationship of the foreground image model fed back by the observer is consistent with the pre-set spectral depth concave-convex relationship of the foreground image model; otherwise, it is inconsistent;

[0061] In step S4, if the observer's spectral depth perception is normal, in N observation evaluations, the concave-convex relationship of the foreground image model fed back by the observer is consistent with the pre-set spectral depth concave-convex relationship of the foreground image model each time.

[0062] For the implementation process of step S2, the pre-set spectral depth concave-convex relationship of the red animal in the foreground image model is based on the concave-convex state observed by an observer with normal spectral depth perception, that is, the pre-set spectral depth concave-convex relationship of the red animal in the foreground image model is based on the "correct" actual situation as the standard.

[0063] In step S5, when setting the display state of the picture content in the display device in each training period, the setting includes: the color of the background canvas, the shape composition of the foreground image model under the color, and the continuous presentation time of the canvas content formed by the combination of the color background canvas and the foreground image model under the color in each training period.

[0064] In the spectral depth perception training of the observer, referring to Figure 6 , the following training steps are included:

[0065] S101. In each training period, the color of the background canvas is set to be unchanged or changed, the picture content is displayed in the display device according to the set continuous presentation time, and the observer feeds back the concave-convex relationship of the foreground image model through the observation equipment.

[0066] In this embodiment, the continuous presentation time can be selected as 20s-60s, and the optimal value is 20s. After the continuous presentation time ends, the color of the background canvas can be changed or not.

[0067] S102. Determine whether the concave-convex relationship of the foreground image model fed back by the observer is consistent with the pre-set spectral depth concave-convex relationship corresponding to the foreground image model. If yes, increase the shape transformation frequency of the foreground image model until the total training period is completed; otherwise, execute step S103.

[0068] In this embodiment, the transformation frequency is selected as 150ms / time-400ms / time, corresponding to the peak value of vp (visual electrophysiology), and the optimal value is 300ms / time. Here, if the concave-convex relationship of the foreground image model fed back by the observer is consistent with the pre-set spectral depth concave-convex relationship corresponding to the foreground image model under the shape transformation frequency of a foreground image model, it means that the observer observes correctly. Therefore, the difficulty of the stimulation condition can be increased. If it is not correct, step S103 is executed.

[0069] S103. Reduce the shape transformation frequency of the foreground image model and return to step S102. If it is not correct, it means that the shape transformation frequency of the original foreground image model is too difficult for the observer participating in the training. Therefore, the difficulty of the stimulation condition can be reduced. Based on this, repeated training can assist in improving the depth perception contour integration function of the observer.

[0070] Embodiment 2

[0071] In this embodiment, unlike in Embodiment 1, the observation device in step S2 is a non-cross-linked 3D spectral separation glasses, such as... Figure 3 As shown, if the background canvas is white, a green circle and a foreground image model of a red animal (elephant) located within the green circle are set on the background canvas. The spectral depth concavity-convexity relationship of the red animal (elephant) in the foreground image model is preset to "convex". If the observer observes the content displayed on the display device through the observation device and reports that the position of the red animal (elephant) is "convex", then the concavity-convexity relationship of the foreground image model reported by the observer is consistent with the preset spectral depth concavity-convexity relationship of the foreground image model; otherwise, they are inconsistent.

[0072] If the background canvas is black, such as Figure 4 As shown, a green circle and a foreground image model of a red animal (elephant) located within the green circle are set on the background canvas. The spectral depth concavity-convexity relationship of the red animal (elephant) in the foreground image model is preset to "concave". If the observer observes the content displayed on the display device through the observation device and reports that the position of the red animal (elephant) is "concave", then the concavity-convexity relationship of the foreground image model reported by the observer is consistent with the preset spectral depth concavity-convexity relationship of the foreground image model; otherwise, they are inconsistent.

[0073] In step S4, if the observer's spectral depth perception is normal, then in N observation evaluations, the concavity and convexity relationship of the foreground image model reported by the observer in each instance is consistent with the spectral depth concavity and convexity relationship corresponding to the pre-set foreground image model.

[0074] Example 3

[0075] like Figure 7 As shown, this embodiment proposes an image-background separation spectral depth perception evaluation and training system, the evaluation and training system comprising:

[0076] The display device is used to display screen content, which includes a background canvas and a foreground image model. With the background canvas, the foreground image model corresponds to different spectral depth concavity and convexity relationships.

[0077] Observation equipment is used by observers to observe the content displayed on a display device. Observers use observation equipment to determine the concavity and convexity relationship of the foreground image model in the content they are viewing.

[0078] The judgment unit is used to determine whether the concavity and convexity relationship of the foreground image model fed back by the observer is consistent with the spectral depth concavity and convexity relationship corresponding to the pre-set foreground image model.

[0079] Obviously, the above embodiments of the present application are only examples for clearly explaining the present application, and are not intended to limit the embodiments of the present application. Based on the above description, other different forms of changes or variations can be made by those skilled in the art. Here, it is not necessary and impossible to exhaust all the embodiments. Any modification, equivalent replacement and improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the claims of the present application.

Claims

1. A spectral depth perception evaluation and training method for image-background separation, characterized in that, The evaluation training methods include: S1. Set the maximum number of times the screen content and observation evaluation are set to N. The screen content includes a background canvas and a foreground image model. With the background canvas, the foreground image model corresponds to different spectral depth concavity and convexity relationships. The background canvas in the image content is a solid color background, a striped background, or a grid background; the width of the background canvas is 2 to 3 times the width of the foreground image model. S2. The screen content is displayed through a display device, and the observer observes the screen content displayed on the display device through an observation device and provides feedback on the concavity and convexity relationship of the foreground image model in the screen content. S3. Determine whether the concavity / convexity relationship of the foreground image model fed back by the observer is consistent with the spectral depth concavity / convexity relationship corresponding to the pre-set foreground image model. If yes, proceed to step S4; otherwise, the observer's spectral depth perception is abnormal, and proceed to step S5. S4. Determine if the number of observations and evaluations has reached N. If yes, the observer's spectral depth perception is normal; otherwise, return to step S2. S5. Set the total training cycle and the display status of the screen content on the display device within each training cycle, and conduct spectral depth perception training for the observer within each training cycle; In step S5, when setting the display status of the screen content on the display device in each training cycle, the following settings are included: the color of the background canvas, the shape composition of the foreground image model under the color, and the continuous presentation time of the canvas content formed by the combination of the background canvas of the color and the foreground image model under the color in each training cycle. Training observers in spectral depth perception includes the following steps: S101. During each training cycle, the background canvas color is set to remain unchanged or change, and the image content is displayed on the display device according to the set continuous presentation time. The observer provides feedback on the concavity and convexity relationship of the foreground image model through the observation device. S102. Determine whether the concavity-convexity relationship of the foreground image model fed back by the observer is consistent with the spectral depth concavity-convexity relationship corresponding to the pre-set foreground image model. If so, increase the morphological transformation frequency of the foreground image model until the total training cycle is completed; otherwise, proceed to step S103. S103. Reduce the morphological transformation frequency of the foreground image model and return to step S102.

2. The image-background separation spectral depth perception evaluation training method according to claim 1, characterized in that, In step S2, when the screen content is displayed through the display device, the background canvas color in the screen content is random, and the foreground image model is also random. However, when the background canvas color changes from color A to color B, the foreground image model matching the background canvas of color A is the same as the foreground image model matching the background canvas of color B.

3. The image-background separation spectral depth perception evaluation training method according to claim 2, characterized in that, In the image content, all components of the foreground image model are static and do not obscure each other.

4. The image-background separation spectral depth perception evaluation training method according to claim 1, characterized in that, The content of the image can be: a foreground image model set on a white background canvas, the foreground image model including a green circle and a red animal located inside the green circle; or it can be: a foreground image model set on a black background canvas, the foreground image model including a green circle and a red animal located inside the green circle; with the combination of different colored background canvases, the red animal located inside the green circle in the foreground image model has a depth and relief relationship with the green circle.

5. The image-background separation spectral depth perception evaluation training method according to claim 4, characterized in that, In step S2, the observation device is a cross 3D spectral separation glasses. If the background canvas is white, a green circle and a foreground image model of a red animal located within the green circle are set on the background canvas. The spectral depth concavity-convexity relationship corresponding to the red animal in the foreground image model is preset to "concave". If the observer observes the content displayed on the display device through the observation device and reports that the position of the red animal is "concave", then the concavity-convexity relationship of the foreground image model reported by the observer is consistent with the spectral depth concavity-convexity relationship corresponding to the preset foreground image model. Otherwise, it is inconsistent; If the background canvas is black, a green circle and a foreground image model of a red animal inside the green circle are set on the background canvas. The spectral depth concavity-convexity relationship of the red animal in the foreground image model is preset to "convex". If the observer observes the content of the screen displayed on the display device through the observation device and reports that the position of the red animal is "convex", then the concavity-convexity relationship of the foreground image model reported by the observer is consistent with the spectral depth concavity-convexity relationship of the foreground image model preset. Otherwise, it is inconsistent; In step S4, if the observer's spectral depth perception is normal, then in N observation evaluations, the concavity and convexity relationship of the foreground image model reported by the observer in each instance is consistent with the spectral depth concavity and convexity relationship corresponding to the pre-set foreground image model.

6. The image-background separation spectral depth perception evaluation training method according to claim 4, characterized in that, In step S2, the observation device is a non-cross-linked 3D spectral separation glasses. If the background canvas is white, a green circle and a foreground image model of a red animal located within the green circle are set on the background canvas. The spectral depth concavity-convexity relationship corresponding to the red animal in the foreground image model is preset to "convex". If the observer observes the content displayed on the display device through the observation device and reports that the position of the red animal is "convex", then the concavity-convexity relationship of the foreground image model reported by the observer is consistent with the spectral depth concavity-convexity relationship corresponding to the preset foreground image model. Otherwise, it is inconsistent; If the background canvas is black, a green circle and a foreground image model of a red animal inside the green circle are set on the background canvas. The spectral depth concavity-convexity relationship of the red animal in the foreground image model is preset to "concave". If the observer observes the content of the screen displayed on the display device through the observation device and reports that the position of the red animal is "concave", then the concavity-convexity relationship of the foreground image model reported by the observer is consistent with the spectral depth concavity-convexity relationship of the foreground image model preset. Otherwise, it is inconsistent; In step S4, if the observer's spectral depth perception is normal, then in N observation evaluations, the concavity and convexity relationship of the foreground image model reported by the observer in each instance is consistent with the spectral depth concavity and convexity relationship corresponding to the pre-set foreground image model.

7. A spectral depth perception evaluation and training system for performing the image-background separation spectral depth perception evaluation and training method of claim 1, characterized in that, The evaluation and training system includes: The display device is used to display screen content, which includes a background canvas and a foreground image model. With the background canvas, the foreground image model corresponds to different spectral depth concavity and convexity relationships. Observation equipment is used by observers to observe the content displayed on a display device. Observers use observation equipment to determine the concavity and convexity relationship of the foreground image model in the content they are viewing. The judgment unit is used to determine whether the concavity and convexity relationship of the foreground image model fed back by the observer is consistent with the spectral depth concavity and convexity relationship corresponding to the pre-set foreground image model.

Images