Visual training method and system based on dichoptic viewing

Abstract

Disclosed in the present invention are a visual training method and system based on dichoptic viewing. The method comprises: determining a dominant eye and a lazy eye of a user according to acquired binocular visual-acuity information of the user, and by means of a dichoptic viewing technique, respectively creating image display channels for the dominant eye and the lazy eye of the user; simultaneously displaying preset first training images in a dominant-eye display channel and a lazy-eye display channel; by means of a preset eye tracking technique and a preset Gaussian blur function, performing Gaussian blurring on each first training image in the dominant-eye display channel, such that by means of the dominant-eye display channel, several second training images which have been subjected to Gaussian blurring are displayed; also displaying the first training images by means of the lazy-eye display channel, such that the visual training of the user is performed; and when a visual training duration of the user reaches a preset time threshold value, completing the visual training of the user. Thus, the effect of visual training is improved.

Description

A vision training method and system based on binocular dichoptic vision Technical Field

[0001] The present invention relates to the technical field of vision training, and in particular to a vision training method and system based on binocular dichoptic vision. Background Art

[0002] The main principle of visual training is to activate the visual system by providing visual stimulation, promote its neural plasticity adjustment, and improve the ability to process visual information.

[0003] Binocular dichoptic technology allows the left and right eyes to see different image information, creating a three-dimensional effect. Each eye receives a different perspective, resulting in the same three-dimensional image.

[0004] Existing visual training methods mainly cover the eye with better vision, or the dominant eye, and perform visual training tasks on the eye with poor vision, or the amblyopic eye (the eye with weaker vision). This enhances the visual stimulation and use of the amblyopic eye, activates visual neurons and pathways, promotes their plastic adjustment, and improves visual information processing capabilities. However, covering one eye has a significant impact on the visual experience, reducing the field of view by half, creating certain obstacles to spatial perception and affecting visual training. At the same time, since only one eye has visual input, the adaptability of the visual environment will be limited, making it impossible to train the coordination and balance of binocular vision functions, thereby reducing the effectiveness of visual training.

[0005] At the same time, existing vision training equipment usually presents dynamic images according to fixed change rules, and cannot dynamically adjust the images according to the user's actual visual conditions. As a result, it is impossible to formulate accurate vision training methods based on the user's actual conditions, which reduces the effect of vision training.

[0006] Summary of the Invention

[0007] In order to solve the above problems, the present invention discloses a vision training method and system based on binocular dichoptic vision to improve the effect of vision training.

[0008] In order to achieve the above objectives, in a first aspect, the present invention discloses a vision training method based on binocular dichoptic vision, comprising:

[0009] Determine the user's dominant eye and amblyopic eye based on the acquired binocular vision information of the user, and create image display channels for the user's dominant eye and amblyopic eye respectively using a preset binocular dichoptic technology; the image display channels include a dominant eye display channel and an amblyopic eye display channel;

[0010] Simultaneously displaying a plurality of preset first training images on the dominant eye display channel and the amblyopic eye display channel, and obtaining the gaze point coordinates of the dominant eye when gazing at the first training image by using a preset eye tracking technology when each of the plurality of first training images is displayed on the dominant eye display channel;

[0011] According to the gaze point coordinates and the binocular vision information, each first training image displayed in the dominant eye display channel is Gaussian blurred using a preset Gaussian blur function to obtain a plurality of second training images after Gaussian blurring;

[0012] displaying the second training image through the dominant eye display channel and simultaneously displaying the first training image through the amblyopic eye display channel to perform vision training on the user, and collecting a duration of the user's vision training;

[0013] When the duration of the visual training reaches a preset time threshold, the playing of the first training image is stopped, and the visual training of the user is completed.

[0014] The present invention discloses a vision training method based on binocular dichopia, which includes obtaining binocular vision information of a user to determine the user's amblyopic eye and dominant eye, so as to perform accurate training according to the user's actual binocular vision conditions, and then using a preset binocular dichopia technology to create image display channels for the user's dominant eye and amblyopic eye respectively, so that the user's left eye and right eye can perform visual training simultaneously, so that both eyes work simultaneously, effectively reducing interocular inhibition, and improving the visual training effect. After creating the image display channel, a preset first image is played in the dominant eye display channel, and the gaze point coordinates of the user when gazing at the first image are obtained by a preset eye tracking technology, and according to the gaze point coordinates and the binocular vision information, a preset Gaussian blur function is used to determine the blurred area for Gaussian blurring in the dominant eye display channel, so that when the preset training images are played in the dominant eye and amblyopic eye display channels at the same time, the training image in the dominant eye display channel is blurred according to the blurred area and the Gaussian blur function, thereby effectively reducing the inhibition of the dominant eye on the amblyopic eye, so that the amblyopic eye of the user receives normal and sharp images during the visual training process, thereby increasing the frequency of use of the amblyopic eye, so as to train the coordination and balance of binocular vision functions and improve the effect of visual training.

[0015] As a preferred example, determining the user's dominant eye and amblyopic eye based on the acquired binocular vision information of the user, and creating image display channels for the user's dominant eye and amblyopic eye respectively through a preset binocular dichoptic technology, includes:

[0016] Obtaining the corrected visual acuity of the user's left eye and right eye respectively, and determining the user's dominant eye and amblyopic eye by comparing the corrected visual acuity of the left eye and the corrected visual acuity of the right eye; the binocular vision information includes the corrected visual acuity of the left eye and the corrected visual acuity of the right eye;

[0017] A first image display channel and a second image display channel are created according to a preset binocular dichoptic technology, so that the user's amblyopic eye and dominant eye respectively gaze at the training image for vision training through the first image display channel and the second image display channel.

[0018] The present invention first obtains the user's binocular vision information to determine the user's dominant eye and amblyopic eye, so as to facilitate the subsequent formulation of a dedicated vision training method based on the user's visual condition and improve the accuracy of vision training. Then, a preset binocular dichoptic technology is used to construct a first image display channel and a second display channel so that the user's two eyes can view the training images for training respectively, ensuring that both eyes are trained simultaneously, thereby improving the coordination and balance of binocular vision function.

[0019] As a preferred example, the plurality of preset first training images are displayed simultaneously on the dominant eye display channel and the amblyopic eye display channel, including:

[0020] Reading a preset training video file, and performing image analysis on the training video file to obtain a plurality of first training images corresponding to the training video file;

[0021] The plurality of first training images are copied to the dominant eye display channel and the amblyopic eye display channel respectively, so that the dominant eye display channel and the amblyopic eye display channel respectively display the plurality of first training images.

[0022] The present invention performs image analysis on the video file used for training and copies the analyzed images to the dominant eye display channel and the amblyopic eye display channel respectively. This first ensures that the user observes a unified image during vision training, ensures the coordinated training of both eyes, and thus improves the effect of vision training.

[0023] As a preferred example, obtaining the gaze point coordinates of the dominant eye when gazing at the first training image by using a preset eye tracking technology includes:

[0024] capturing, using a preset eye tracking sensor, an eye tracking sight line of the user's dominant eye when the user is looking at the first training image;

[0025] The coordinates of the intersection of the eye tracking line of sight and the first training image are calculated by using a coordinate calculation function preset in the eye tracking sensor to obtain the coordinates of the gaze point where the eye tracking line of sight of the user's dominant eye falls on the first training image.

[0026] When the dominant eye display channel plays the first training image, the present invention uses a preset eye tracking technology to capture the eye tracking line of the user's dominant eye looking at the first training image in real time to obtain the gaze point coordinates of the dominant eye falling on the first training image, so as to facilitate the subsequent determination of the Gaussian blur area based on the gaze point coordinates, thereby blurring the training image in the display channel where the dominant eye is located, thereby increasing the frequency of the user using the amblyopic eye and achieving visual training.

[0027] As a preferred example, performing Gaussian blur on each first training image displayed in the dominant eye display channel by using a preset Gaussian blur function according to the gaze point coordinates and the binocular vision information includes:

[0028] Determining the convolution kernel value, the two-dimensional direction standard deviation, and the radius of the blurred area of ​​the fixation point in the Gaussian blur function according to the corrected visual acuity of the amblyopic eye;

[0029] Obtaining a blurred area of ​​each first training image displayed in the dominant eye display channel with the gaze point coordinates as the center according to the gaze point blurred area radius;

[0030] According to the convolution kernel value and the two-dimensional standard deviation, Gaussian blur is performed on the blurred area of ​​each first training image using the Gaussian blur function to obtain a Gaussian blurred second training image.

[0031] The present invention utilizes the corrected vision of the amblyopic eye to determine the parameter values ​​in the Gaussian blur function, that is, to ensure that the image in the dominant eye channel is Gaussian blurred according to the amblyopia degree of the amblyopic eye. Different blurring degrees are determined for different amblyopia degrees, so that different visual training degrees are specified for the visual conditions of different users, better meeting the visual training needs of different users. At the same time, the corrected vision of the amblyopic eye is utilized to blur the image in the dominant eye display channel, so that the dominant eye receives an image with regional blurring, while the amblyopic eye receives a normal, sharp image. This effectively reduces the inhibition of the dominant eye on the amblyopic eye, increases the frequency of users using the amblyopic eye, and thereby improves the visual training effect of the users.

[0032] In a second aspect, the present invention discloses a vision training system based on binocular dichoptic vision, comprising an image display module, an eye tracking module, an image processing module, a vision training module, and a training end module;

[0033] The image display module is used to determine the user's dominant eye and amblyopic eye based on the acquired binocular vision information of the user, and to create image display channels for the user's dominant eye and amblyopic eye respectively through a preset binocular dichoptic technology; the image display channels include a dominant eye display channel and an amblyopic eye display channel;

[0034] The eye tracking module is configured to simultaneously display a plurality of preset first training images on the dominant eye display channel and the amblyopic eye display channel, and obtain, by using a preset eye tracking technology, the gaze point coordinates of the dominant eye when gazing at the first training image when each of the plurality of first training images is displayed on the dominant eye display channel;

[0035] The image processing module is used to perform Gaussian blurring on each first training image displayed in the dominant eye display channel using a preset Gaussian blur function according to the gaze point coordinates and the binocular vision information, to obtain a plurality of second training images after Gaussian blurring;

[0036] The vision training module is configured to display the second training image through the dominant eye display channel and simultaneously display the first training image through the amblyopic eye display channel to perform vision training on the user, and collect a duration of the user's vision training;

[0037] The training end module is used to stop playing the first training image when the visual training duration reaches a preset time threshold, thereby completing the user's visual training.

[0038] The present invention discloses a vision training system based on binocular dichoptic vision, which includes obtaining binocular vision information of a user to determine the user's amblyopic eye and dominant eye, so as to perform accurate training according to the user's actual binocular vision conditions, and then using a preset binocular dichoptic vision technology to create image display channels for the user's dominant eye and amblyopic eye respectively, so that the user's left eye and right eye can perform visual training simultaneously, so that both eyes work simultaneously, effectively reducing interocular inhibition, and improving the visual training effect. After creating the image display channel, a preset first image is played in the dominant eye display channel, and the gaze point coordinates of the user when gazing at the first image are obtained by a preset eye tracking technology, and according to the gaze point coordinates and the binocular vision information, a preset Gaussian blur function is used to determine the blurred area for Gaussian blurring in the dominant eye display channel, so that when the preset training images are played in the dominant eye and amblyopic eye display channels at the same time, the training image in the dominant eye display channel is blurred according to the blurred area and the Gaussian blur function, thereby effectively reducing the inhibition of the dominant eye on the amblyopic eye, so that the amblyopic eye of the user receives normal and sharp images during the visual training process, thereby increasing the frequency of use of the amblyopic eye, so as to train the coordination and balance of binocular vision functions and improve the effect of visual training.

[0039] As a preferred example, the image display module includes an information acquisition unit and a channel display unit;

[0040] The information acquisition unit is used to respectively acquire the user's left eye corrected visual acuity and right eye corrected visual acuity, and determine the user's dominant eye and amblyopic eye by comparing the left eye corrected visual acuity and the right eye corrected visual acuity; the binocular vision information includes the left eye corrected visual acuity and the right eye corrected visual acuity;

[0041] The channel display unit is used to create a first image display channel and a second image display channel according to a preset binocular dichoptic technology, so that the user's amblyopic eye and dominant eye can respectively gaze at the training image for vision training through the first image display channel and the second image display channel.

[0042] The present invention first obtains the user's binocular vision information to determine the user's dominant eye and amblyopic eye, so as to facilitate the subsequent formulation of a dedicated vision training method based on the user's visual condition and improve the accuracy of vision training. Then, a preset binocular dichoptic technology is used to construct a first image display channel and a second display channel so that the user's two eyes can view the training images for training respectively, ensuring that both eyes are trained simultaneously, thereby improving the coordination and balance of binocular vision function.

[0043] As a preferred example, the eye tracking module includes an image display unit and an eye tracking unit;

[0044] The image display unit is used to read a preset training video file and perform image analysis on the training video file to obtain a plurality of first training images corresponding to the training video file; and copy the plurality of first training images to the dominant eye display channel and the amblyopic eye display channel, respectively, so that the dominant eye display channel and the amblyopic eye display channel respectively display the plurality of first training images;

[0045] The eye tracking unit is used to capture the eye tracking line of sight of the user's dominant eye when it is looking at the first training image according to a preset eye tracking sensor; calculate the coordinates of the intersection of the eye tracking line of sight and the first training image using a coordinate calculation function preset in the eye tracking sensor, and obtain the coordinates of the gaze point where the eye tracking line of sight of the user's dominant eye falls on the first training image.

[0046] The present invention performs image analysis on a video file used for training and copies several analyzed images to the dominant eye display channel and the amblyopic eye display channel respectively. First, it ensures that the user observes a unified image during visual training, ensures the coordinated training of both eyes, and thus improves the effect of visual training. Then, when the dominant eye display channel plays a first training image, a preset eye tracking technology is used to capture the eye tracking sight of the user's dominant eye looking at the first training image in real time to obtain the coordinates of the gaze point of the dominant eye falling on the first training image, so as to facilitate the subsequent determination of the Gaussian blur area according to the gaze point coordinates, and blur the training image in the display channel where the dominant eye is located, thereby increasing the frequency of the user using the amblyopic eye and thus achieving visual training.

[0047] As a preferred example, the image processing module includes a parameter determination unit, an area calculation unit and a Gaussian blur unit;

[0048] The parameter determination unit is used to determine the convolution kernel value, the two-dimensional direction standard deviation and the radius of the blurred area of ​​the fixation point in the Gaussian blur function according to the corrected visual acuity of the amblyopic eye;

[0049] The area calculation unit is used to obtain a blurred area of ​​each first training image displayed in the dominant eye display channel based on the blurred area radius of the gaze point and with the gaze point coordinates as the center;

[0050] The Gaussian blur unit is used to perform Gaussian blur on the blurred area of ​​each first training image according to the convolution kernel value and the two-dimensional standard deviation through the Gaussian blur function to obtain a Gaussian blurred second training image.

[0051] The present invention utilizes the corrected vision of the amblyopic eye to determine the parameter values ​​in the Gaussian blur function, that is, to ensure that the image in the dominant eye channel is Gaussian blurred according to the amblyopia degree of the amblyopic eye. Different blurring degrees are determined for different amblyopia degrees, so that different visual training degrees are specified for the visual conditions of different users, better meeting the visual training needs of different users. At the same time, the corrected vision of the amblyopic eye is utilized to blur the image in the dominant eye display channel, so that the dominant eye receives an image with regional blurring, while the amblyopic eye receives a normal, sharp image. This effectively reduces the inhibition of the dominant eye on the amblyopic eye, increases the frequency of users using the amblyopic eye, and thereby improves the visual training effect of the users.

[0052] In a third aspect, the present invention discloses a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a visual training method based on binocular dichoptic vision as described in the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] FIG1 is a flow chart of a vision training method based on binocular dichoptic viewing provided by an embodiment of the present invention;

[0054] FIG2 is a schematic structural diagram of a vision training system based on binocular dichoptic viewing provided by an embodiment of the present invention;

[0055] FIG3 is a flow chart of a vision training method based on binocular dichoptic viewing according to another embodiment of the present invention. DETAILED DESCRIPTION

[0056] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0057] Example 1

[0058] An embodiment of the present invention provides a vision training method based on binocular dichoptic viewing. The specific implementation process of the method is shown in FIG1 , which mainly includes steps 101 to 105. The steps mainly include:

[0059] Step 101: Determine the user's dominant eye and amblyopic eye based on the acquired binocular vision information of the user, and create image display channels for the user's dominant eye and amblyopic eye respectively through a preset binocular dichoptic technology; the image display channels include a dominant eye display channel and an amblyopic eye display channel.

[0060] In this embodiment, this step mainly includes: obtaining the user's left eye corrected visual acuity and right eye corrected visual acuity respectively, and determining the user's dominant eye and amblyopic eye by comparing the left eye corrected visual acuity and the right eye corrected visual acuity; the binocular vision information includes the left eye corrected visual acuity and the right eye corrected visual acuity; creating a first image display channel and a second image display channel according to a preset binocular dichoptic technology, so that the user's amblyopic eye and dominant eye respectively look at the training image for visual training through the first image display channel and the second image display channel.

[0061] This embodiment first obtains the user's binocular vision information to determine the user's dominant eye and amblyopic eye, so that a customized vision training method can be developed based on the user's visual condition in the future to improve the accuracy of vision training. Then, a preset binocular dichoptic technology is used to construct a first image display channel and a second display channel so that the user's two eyes can view the training images for training respectively, ensuring that both eyes are trained simultaneously, thereby improving the coordination and balance of binocular vision function.

[0062] Step 102: Simultaneously displaying a plurality of preset first training images on the dominant eye display channel and the amblyopic eye display channel, and obtaining the gaze point coordinates of the dominant eye when gazing at the first training image by using a preset eye tracking technology when each of the plurality of first training images is displayed on the dominant eye display channel.

[0063] In this embodiment, this step mainly includes: reading a preset training video file and performing image analysis on the training video file to obtain a plurality of first training images corresponding to the training video file; copying the plurality of first training images to the dominant eye display channel and the amblyopic eye display channel, respectively, so that the dominant eye display channel and the amblyopic eye display channel respectively display the plurality of first training images; capturing the eye tracking line of sight of the user's dominant eye when it is looking at the first training image using a preset eye tracking sensor; calculating the coordinates of the intersection of the eye tracking line of sight and the first training image using a coordinate calculation function preset in the eye tracking sensor, thereby obtaining the coordinates of the gaze point where the eye tracking line of sight of the user's dominant eye falls on the first training image.

[0064] This embodiment performs image analysis on a video file used for training and copies the analyzed images to the dominant eye display channel and the amblyopic eye display channel, respectively. This ensures that the user observes a unified image during visual training, ensuring binocular coordination training and thereby improving the effectiveness of visual training. Then, while the dominant eye display channel plays a first training image, the user's dominant eye's eye tracking line of sight as it gazes at the first training image is captured in real time using a preset eye-tracking technology. This obtains the coordinates of the dominant eye's gaze point on the first training image, facilitating the subsequent determination of the Gaussian blur region based on the gaze point coordinates. This blurs the training image in the dominant eye's display channel, thereby increasing the user's frequency of using the amblyopic eye and achieving visual training.

[0065] Step 103: Based on the gaze point coordinates and the binocular vision information, Gaussian blur is performed on each first training image displayed in the dominant eye display channel using a preset Gaussian blur function to obtain a plurality of second training images after Gaussian blur.

[0066] In this embodiment, this step mainly includes: determining the convolution kernel value, two-dimensional standard deviation and gaze point blurred area radius in the Gaussian blur function according to the corrected visual acuity of the amblyopic eye; obtaining the blurred area of ​​each first training image displayed in the dominant eye display channel with the gaze point coordinates as the center according to the gaze point blurred area radius; and performing Gaussian blur on the blurred area of ​​each first training image according to the convolution kernel value and the two-dimensional standard deviation through the Gaussian blur function to obtain a second training image after Gaussian blur.

[0067] This embodiment uses the corrected vision of the amblyopic eye to determine the parameter values ​​in the Gaussian blur function, that is, to ensure that the image in the dominant eye channel is Gaussian blurred according to the amblyopia degree of the amblyopic eye. Different degrees of blur are determined for different amblyopia degrees, so that different visual training degrees are specified for the visual conditions of different users, and the visual training needs of different users are better met. At the same time, the corrected vision of the amblyopic eye is used to blur the image in the dominant eye display channel, so that the dominant eye receives the blurred image in the area, while the amblyopic eye receives the normal, sharp image. This effectively reduces the inhibition of the dominant eye on the amblyopic eye, increases the frequency of users using the amblyopic eye, and thereby improves the user's visual training effect.

[0068] Step 104: Display the second training image through the dominant eye display channel, and simultaneously display the first training image through the amblyopic eye display channel to perform vision training on the user, and collect the user's vision training duration.

[0069] In this embodiment, this step mainly includes: using the second training image after Gaussian blur to promptly replace the first training image played in the dominant eye display channel, so that the dominant eye display channel displays the second training image while the amblyopic eye display channel displays the preset first training image, thereby performing visual training for the user's two eyes and collecting the user's visual training time.

[0070] This embodiment simultaneously displays a Gaussian blurred image in the dominant eye and a clear original image in the amblyopic eye, thereby causing the user's eyes to view different images simultaneously, unconsciously increasing the frequency of use of the amblyopic eye, thereby forming visual training for the amblyopic eye and improving the effect of visual training.

[0071] Step 105: When the visual training duration reaches a preset time threshold, the playback of the first training image is stopped, and the visual training of the user is completed.

[0072] In this embodiment, this step mainly includes: when the duration of the user's visual training exceeds a preset time threshold, stopping the playback of the first training image to complete the user's visual training.

[0073] The present invention completes each visual training of the user by collecting the visual training duration of the user, so as to allow the user's eyes to rest and protect the effect of the visual training.

[0074] In this embodiment, a visual training system based on binocular dichoptic viewing is also provided. Please refer to Figure 2 for the specific structural composition of the system, which mainly includes an image display module 201, an eye tracking module 202, an image processing module 203, a visual training module 204 and a training end module 205.

[0075] The image display module 201 is used to determine the user's dominant eye and amblyopic eye based on the acquired binocular vision information of the user, and to create image display channels for the user's dominant eye and amblyopic eye respectively through a preset binocular dichoptic technology; the image display channels include a dominant eye display channel and an amblyopic eye display channel.

[0076] The eye tracking module 202 is used to simultaneously display several preset first training images on the dominant eye display channel and the amblyopic eye display channel, and when each of the several first training images is displayed on the dominant eye display channel, obtain the gaze point coordinates of the dominant eye when it looks at the first training image through a preset eye tracking technology.

[0077] The image processing module 203 is used to perform Gaussian blur on each first training image displayed in the dominant eye display channel according to the gaze point coordinates and the binocular vision information using a preset Gaussian blur function to obtain a plurality of second training images after Gaussian blur.

[0078] The vision training module 204 is configured to display the second training image through the dominant eye display channel and simultaneously display the first training image through the amblyopic eye display channel to perform vision training on the user and collect a duration of the user's vision training.

[0079] The training end module 205 is configured to stop playing the first training image when the visual training duration reaches a preset time threshold, thereby completing the user's visual training.

[0080] In this embodiment, the image display module 201 includes an information acquisition unit and a channel display unit.

[0081] The information acquisition unit is used to respectively acquire the user's left eye corrected vision and right eye corrected vision, and determine the user's dominant eye and amblyopia by comparing the left eye corrected vision and the right eye corrected vision. The binocular vision information includes the left eye corrected vision and the right eye corrected vision.

[0082] The channel display unit is used to create a first image display channel and a second image display channel according to a preset binocular dichoptic technology, so that the user's amblyopic eye and dominant eye can respectively gaze at the training image for vision training through the first image display channel and the second image display channel.

[0083] In this embodiment, the eye tracking module 202 includes an image display unit and an eye tracking unit.

[0084] The image display unit is used to read a preset training video file and perform image analysis on the training video file to obtain a plurality of first training images corresponding to the training video file; and copy the plurality of first training images to the dominant eye display channel and the amblyopic eye display channel, respectively, so that the dominant eye display channel and the amblyopic eye display channel respectively display the plurality of first training images.

[0085] The eye tracking unit is used to capture the eye tracking line of sight of the user's dominant eye when looking at the first training image according to a preset eye tracking sensor; calculate the coordinates of the intersection of the eye tracking line of sight and the first training image by using a coordinate calculation function preset in the eye tracking sensor, and thereby obtain the coordinates of the gaze point where the eye tracking line of sight of the user's dominant eye falls on the first training image.

[0086] In this embodiment, the image processing module 203 includes a parameter determination unit, a region calculation unit, and a Gaussian blur unit.

[0087] The parameter determination unit is used to determine the convolution kernel value, the two-dimensional direction standard deviation and the radius of the gaze point blur area in the Gaussian blur function according to the corrected vision of the amblyopic eye.

[0088] The area calculation unit is used to obtain the blurred area of ​​each first training image displayed in the dominant eye display channel based on the blurred area radius of the gaze point and with the gaze point coordinates as the center.

[0089] The Gaussian blur unit is used to perform Gaussian blur on the blurred area of ​​each first training image according to the convolution kernel value and the two-dimensional standard deviation through the Gaussian blur function to obtain a Gaussian blurred second training image.

[0090] In addition to the above-mentioned method and system, an embodiment of the present invention further discloses a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the visual training method based on binocular vision as described in an embodiment of the present invention is implemented.

[0091] The present invention discloses a vision training method and system based on binocular dichopia, which obtains the user's binocular vision information to determine the user's amblyopic eye and dominant eye, so as to perform precise training according to the user's actual binocular vision conditions. Then, a preset binocular dichopia technology is used to create image display channels for the user's dominant eye and amblyopic eye respectively, so that the user's left eye and right eye can perform visual training simultaneously, so that both eyes work simultaneously, effectively reducing interocular inhibition and improving the visual training effect. After creating the image display channel, a preset first image is played in the dominant eye display channel, and the gaze point coordinates of the user when gazing at the first image are obtained by a preset eye tracking technology, and according to the gaze point coordinates and the binocular vision information, a preset Gaussian blur function is used to determine the blurred area for Gaussian blurring in the dominant eye display channel, so that when the preset training images are played in the dominant eye and amblyopic eye display channels at the same time, the training image in the dominant eye display channel is blurred according to the blurred area and the Gaussian blur function, thereby effectively reducing the inhibition of the dominant eye on the amblyopic eye, so that the amblyopic eye of the user receives normal and sharp images during the visual training process, thereby increasing the frequency of use of the amblyopic eye, so as to train the coordination and balance of binocular vision functions and improve the effect of visual training.

[0092] Example 2

[0093] In an embodiment of the present invention, another vision training method based on binocular dichoptic viewing is provided. The specific implementation process of the method is shown in FIG3 , which mainly includes steps 301 to 304. The steps mainly include:

[0094] Step 301: Obtain the user's binocular vision information to determine the amblyopic eye and the dominant eye, and construct the user's dominant eye image display channel and the amblyopic eye image display channel through a preset binocular dichoptic technology, and simultaneously play the preset first training image through the dominant eye image display channel and the amblyopic eye image display channel respectively.

[0095] In this embodiment, this step mainly includes: determining the user's dominant eye and amblyopic eye based on the acquired binocular vision information of the user, and creating image display channels for the user's dominant eye and amblyopic eye respectively through a preset binocular dichoptic technology, reading a preset training video file, and performing image analysis on the training video file to obtain a number of first training images corresponding to the training video file; and copying the several first training images to the dominant eye display channel and the amblyopic eye display channel, respectively, so that the dominant eye display channel and the amblyopic eye display channel respectively display the several first training images.

[0096] In this embodiment, this step is specifically as follows: obtaining basic information of the user's binocular vision, including the user's left and right eye corrected vision, gender, and age. In this embodiment, the user's right eye corrected vision is set to 0.6, and the left eye corrected vision is set to 0.3. Then, through the corrected vision, it can be determined that the user's right eye is the dominant eye and the left eye is the amblyopic eye. Furthermore, if the vision of both eyes is the same, no processing will be performed. In this embodiment, before conducting visual training on the user, the user is asked to sit 50 cm in front of the display screen. Then, the preset binocular dichoptic technology is used to create image display channels for the user's left and right eyes respectively, so that the user's left and right eyes can respectively view the training images for visual training through the image display channels.

[0097] In this embodiment, the binocular dichotomy technology includes a series of technologies that can achieve binocular dichotomy, such as naked-eye 3D devices, polarized 3D displays, or binocular dichotomy through color filter glasses. In this embodiment, when using a naked-eye 3D device, two independent display channels are created, named the right-eye display channel right_channel and the left-eye display channel left_channel, to adjust the content for each eye separately. left_channel = np.zeros(frame.shape, dtype = np.uint8); right_channel = np.zeros(frame.shape, dtype = np.uint8). Binocular dichotomy can also be achieved using color filter glasses. In this embodiment, the color filter glasses can use red-blue glasses, with blue lenses for the left eye and red lenses for the right eye. The different lenses allow the user's left and right eyes to view the training images separately.

[0098] After creating the left and right eye display channels, open the video file cap = cv2.VideoCapture('video.mp4'). cap is a VideoCapture object, which can be used to read the video file. Each image frame can be processed in real time within a while loop, reading a video frame ret , frame = cap.read() . Each video frame is a numpy array, so the variable frame can be used directly to represent the image. Next, copy the displayed image to both display channels. Copy the image to the left eye display channel with left_channel[:] = frame, and copy the image to the right eye display channel with right_channel[:] = frame. Furthermore, when using the green eye for binocular dichroism, use the split function in OpenCV to split the original training image into three channels (RGB). Operate on the channel values ​​so that the right eye sees only the R channel and the left eye sees only the G and B channels. Finally, use the merge function to merge the processed channels into a single image. Before merging the left and right eye channels into the final image, adjust the parallax effect by appropriately offsetting the two processed images. You can use the warpAffine function in OpenCV to perform operations such as image translation. The offset can be adjusted to suit your needs. The processed left and right eye channels are used as inputs to the merge function, and the output image is designated as the final composite image, which is then displayed.

[0099] Step 302: Obtain the gaze point coordinates of the dominant eye when it is gazing at the first training image according to a preset eye tracking technology, and perform Gaussian blur on the first training image in the display channel where the dominant eye is located using a preset Gaussian blur function according to the gaze point coordinates, so that the display channel where the dominant eye is located plays the second training image after Gaussian blur.

[0100] In this embodiment, this step mainly includes: obtaining the gaze point coordinates of the user's dominant eye falling on the first training image according to a preset eye tracking sensor, and then Gaussian blurring the first training image in the dominant eye display channel using a preset Gaussian blur function according to the gaze point coordinates to obtain a second training image after Gaussian blurring.

[0101] In this embodiment, the specific steps are as follows: For example, in this embodiment, a preset eye tracking sensor is used to capture gaze, and eye tracking can also be performed by using a gaze capture function or other eye-realization technology. In this embodiment, eye tracking is performed based on a Tobii eye tracking sensor. In the TobiiSDK, an eye tracking sensor connection is created through tracker = tobii_research.find_all_eyetrackers()[0]; tracker.connect(), and then the gaze point coordinates of the right eye at the current moment are obtained through gaze_data = tracker.get_gaze_data(); right_gaze_point = gaze_data["right_gaze_point_on_display_area"]. The obtained right eye gaze point coordinate value is relative to the viewport coordinate system and can be directly used to render a graphical interface or perform interface-related calculations and other operations.

[0102] In this embodiment, it is known that the right eye display channel is the dominant eye display channel, and the right eye display channel image needs to be processed. With the right eye's gaze point coordinates as right_gaze_point as the center, the area to be blurred is calculated and controlled in real time and Gaussian blur is performed. In this embodiment, the GaussianBlur function of OpenCV can be used for Gaussian blur. The parameters of the Gaussian blur function GaussianBlur() have the following meanings: cv2.GaussianBlur(src,ksize,sigmaX[,dst[,sigmaY[,borderType]]])→dst

[0103] Where src represents the original image to be Gaussian blurred, ksize represents the kernel size, expressed as (width, height), and must be positive and odd. If only a positive integer value ksize is given, the kernel size is assumed to be (ksize, ksize). sigmaX represents the standard deviation in the X direction; a larger value results in smaller values ​​within the kernel, resulting in less detail in the image and greater blurring. dst represents the destination image, an output parameter of the function, which defaults to None. sigmaY represents the standard deviation in the Y direction. If not specified, it defaults to the same as sigmaX. borderType represents the border mode used to infer pixels outside the image. The default is cv2.BORDER_DEFAULT.

[0104] When ksize is set to a smaller value, the blur is less severe, resulting in more image detail. When ksize is set to a larger value, the blur is greater, resulting in less image detail. If ksize is set too large, not only will the computational effort increase, but it may also cause image distortion. The sigmaX and sigmaY parameters control the standard deviation of the Gaussian kernel function in the X and Y directions, respectively. Larger values ​​result in smaller values ​​within the convolution kernel, resulting in less image detail, which intuitively translates to greater blur. If sigmaX and sigmaY have different values, a non-circular kernel will be generated. borderType is the border mode inferred for pixels outside the image. The default value, BORDER_DEFAULT, is typically used. Using complex zero-padding methods for image borders may affect blur performance and image quality. This selection and adjustment should be tailored to the specific application scenario.

[0105] In this embodiment, different convolution kernel sizes, standard deviations, and radius of the blurred area of ​​the dominant eye's gaze point are set according to the different degrees of amblyopia of the amblyopic eye to achieve different degrees of Gaussian blur. The more severe the amblyopia, the greater the degree of blur required. At the same time, an interface is provided so that the trainer can adjust parameters according to the user's actual situation or based on user feedback, including blur degree, blur radius, etc. The system will adjust the parameters according to the trainer's settings. The parameters exemplified in the present invention are only examples for the convenience of presenting the technical solution.

[0106] In this embodiment, amblyopia can be divided into severe, moderate and mild: visual acuity below 0.2 is severe amblyopia, 0.2-0.5 is moderate amblyopia, and 0.6-0.8 is mild amblyopia.

[0107] When the amblyopia is mild, the blurring effect is lighter, mainly removing the central details, while the peripheral contours are still visible. The convolution kernel is ksize(3,3), sigma=1, blur_rect_size(100,100);

[0108] When the amblyopia is moderate, the center is obviously blurred and the periphery also begins to blur, but the whole image is still recognizable. The convolution kernel is ksize(5,5), sigma=1.5, blur_rect_size(150,150);

[0109] When the amblyopia is severe, the image appears blurry, with only rough colors and outlines distinguishable. The convolution kernel is ksize(7,7), sigma=2, and blur_rect_size(200,200). The rectangle of the blurred area is calculated based on the gaze point coordinates.

[0110] blur_rect=((right_gaze_point.x-blur_rect_size[0] / / 2,right_gaze_point.y-blur_rect_size[1] / / 2),(right_gaze_point.x+blur_rect_size[0] / / 2,right_gaze_point.y+blur_rect_size[1] / / 2))

[0111] Read the right-eye display channel image right_img and perform Gaussian blur on the blurred area.

[0112] right_img[blur_rect[0][1]:blur_rect[1][1],blur_rect[0][0]:blur_rect[1][0]]=cv2.GaussianBlur(right_img[blur_rect[0][1]:blur_rect[1][1],blur_rect[0][0]:blur_rect[1][0]],ksize,sigma)

[0113] After the Gaussian blur is completed, the right eye display channel is updated to the processed image. The dominant eye display channel is blurred only in the fixation area while keeping the rest of the image clear. The amblyopic eye display channel is a normal, sharp image.

[0114] Step 303: The second training image is displayed through the dominant eye image display channel while the first training image is displayed through the amblyopic eye display channel to perform vision training on the user. When the duration of the user's vision training is greater than a preset time threshold, the user's vision training is terminated.

[0115] In this embodiment, the specific step is: while playing the blurred second training image to the user's dominant eye, the original training image without Gaussian blurring corresponding to the second training image is played to the user's amblyopic eye. The user is undergoing visual training while watching the video, and the compliance is very good. Each training session can be set to watch for 15-30 minutes, allowing the user to improve their vision and restore their visual function unknowingly while watching their favorite videos.

[0116] In this embodiment, during training, the user's eye movement data, including eye movement trajectories, can be recorded for analysis of user compliance. The trainer should regularly check the user's vision improvement and adjust the blur parameters accordingly, dynamically updating and adjusting the parameters to adapt to the improved visual state.

[0117] The present invention discloses a binocular vision training method based on dichoptic vision. This method, powered by an eye-tracking sensor and an innovative, real-time image processing algorithm, divides visual content into two independent channels, allowing for separate manipulation of the content for each eye. The system blurs the center of vision of the dominant eye based on the severity of amblyopia, while maintaining clarity in the rest of the image. This blurring is performed in real time based on the dominant eye's instantaneous gaze position, while the amblyopic eye receives a normal, sharp image. This effectively reduces interocular inhibition, allowing for greater utilization of the amblyopic eye and increasing its frequency of use, allowing users to unconsciously improve their vision and restore visual function while viewing the screen.

[0118] Compared to traditional computer vision training, the method provided in this embodiment does not require the development of specialized visual training tasks. Instead, it can leverage numerous existing multimedia resources. For example, a user can open a video or video website and watch their favorite movie or cartoon. The system will process the images in real time while the user is watching the video, allowing the user to complete visual training unconsciously. This can greatly improve user compliance and does not require the user to use interactive operations such as a mouse. Compared with existing technical solutions, this invention is simpler to use and has better compliance.

[0119] The specific embodiments described above further illustrate the objectives, technical solutions, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A visual training method based on binocular dichoptic, characterized in that: include: Determine the user's dominant eye and amblyopic eye according to the acquired binocular vision information of the user, and create image display channels for the dominant eye and amblyopic eye of the user respectively through a preset binocular dichoptic technology; the image display channels include a dominant eye display channel and an amblyopic eye display channel; The dominant eye display channel and the amblyopic eye display channel simultaneously display a plurality of preset first training images, and when the dominant eye display channel displays each of the plurality of first training images, the gaze point coordinates of the dominant eye when gazing at the first training image are obtained by using a preset eye tracking technology; According to the gaze point coordinates and the binocular vision information, each first training image displayed in the dominant eye display channel is Gaussian blurred by a preset Gaussian blur function to obtain a plurality of second training images after Gaussian blur; displaying the second training image through the dominant eye display channel and displaying the first training image through the amblyopic eye display channel to perform visual training for the user, and collecting the visual training duration of the user; When the duration of the visual training reaches a preset time threshold, the playing of the first training image is stopped, and the visual training of the user is completed.

2. A visual training method based on binocular dichoptics as claimed in claim 1, characterized in that: The method of determining the dominant eye and the amblyopic eye of the user according to the acquired binocular vision information of the user, and creating image display channels for the dominant eye and the amblyopic eye of the user respectively by using a preset binocular dichoptic technology, includes: Respectively obtaining the corrected vision of the left eye and the corrected vision of the right eye of the user, and determining the dominant eye and the amblyopic eye of the user by comparing the corrected vision of the left eye and the corrected vision of the right eye; the binocular vision information includes the corrected vision of the left eye and the corrected vision of the right eye; A first image display channel and a second image display channel are created according to a preset binocular dichoptic technology, so that the user's amblyopic eye and dominant eye respectively gaze at a training image for visual training through the first image display channel and the second image display channel.

3. A visual training method based on binocular dichoptic as claimed in claim 1, characterized in that: The step of simultaneously displaying a plurality of preset first training images on the dominant eye display channel and the amblyopic eye display channel includes: Reading a preset training video file, and performing image analysis on the training video file to obtain a plurality of first training images corresponding to the training video file; The plurality of first training images are copied to the dominant eye display channel and the amblyopic eye display channel respectively, so that the dominant eye display channel and the amblyopic eye display channel display the plurality of first training images respectively.

4. A method for visual training based on binocular dichoptics as claimed in claim 1, characterized in that: The obtaining the gaze point coordinates of the dominant eye when gazing at the first training image by using a preset eye tracking technology includes: capturing the eye tracking sight of the user's dominant eye when the user is gazing at the first training image according to a preset eye tracking sensor; The coordinate calculation function preset in the eye tracking sensor calculates the coordinates of the eye tracking sight and the first training image. The coordinates of the intersection of the eye tracking sight of the dominant eye of the user falling on the first training image are obtained.

5. A visual training method based on binocular dichoptic as claimed in claim 2, characterized in that: The step of performing Gaussian blurring on each first training image displayed in the dominant eye display channel by using a preset Gaussian blur function according to the gaze point coordinates and the binocular vision information includes: Determining the convolution kernel value, the two-dimensional direction standard deviation and the radius of the blurred area of ​​the fixation point in the Gaussian blur function according to the corrected vision of the amblyopic eye; According to the radius of the fuzzy area of ​​the gaze point, a fuzzy area of ​​each first training image displayed in the dominant eye display channel is obtained with the coordinates of the gaze point as the center; According to the convolution kernel value and the two-dimensional standard deviation, Gaussian blur is performed on the blurred area of ​​each first training image by using the Gaussian blur function to obtain a second training image after Gaussian blur.

6. A vision training system based on binocular dichoptic, characterized in that: It includes an image display module, an eye tracking module, an image processing module, a visual training module and a training end module; The image display module is used to determine the dominant eye and amblyopic eye of the user according to the acquired binocular vision information of the user, and to create image display channels for the dominant eye and amblyopic eye of the user respectively through a preset binocular dichoptic technology; the image display channels include a dominant eye display channel and an amblyopic eye display channel; The eye tracking module is used to simultaneously display a plurality of preset first training images on the dominant eye display channel and the amblyopic eye display channel, and when each of the plurality of first training images is displayed on the dominant eye display channel, obtain the gaze point coordinates of the dominant eye when gazing at the first training image by using a preset eye tracking technology; The image processing module is used to perform Gaussian blurring on each first training image displayed in the dominant eye display channel according to the gaze point coordinates and the binocular vision information by using a preset Gaussian blur function to obtain a plurality of second training images after Gaussian blurring; The visual training module is used to display the second training image through the dominant eye display channel and display the first training image through the amblyopic eye display channel to perform visual training for the user, and collect the visual training duration of the user; The training end module is used to stop playing the first training image when the duration of the visual training reaches a preset time threshold, thereby completing the visual training of the user.

7. A vision training system based on binocular dichoptic viewing as claimed in claim 6, characterized in that: The image display module includes an information acquisition unit and a channel display unit; The information acquisition unit is used to respectively acquire the user's left eye corrected vision and right eye corrected vision, and determine the user's dominant eye and amblyopic eye by comparing the left eye corrected vision and the right eye corrected vision; the binocular vision information includes the left eye corrected vision and the right eye corrected vision; The channel display unit is used to create a first image display channel and a second image display channel according to a preset binocular dichoptic technology, so that the user's amblyopic eye and dominant eye can respectively watch the image for visual training through the first image display channel and the second image display channel. Training images.

8. A vision training system based on binocular dichoptic viewing as claimed in claim 6, characterized in that: The eye tracking module includes an image display unit and an eye tracking unit; The image display unit is used to read a preset training video file, and perform image analysis on the training video file to obtain a plurality of first training images corresponding to the training video file; and copy the plurality of first training images to the dominant eye display channel and the amblyopic eye display channel, respectively, so that the dominant eye display channel and the amblyopic eye display channel display the plurality of first training images, respectively; The eye tracking unit is used to capture the eye tracking sight line of the user's dominant eye when the user is gazing at the first training image according to a preset eye tracking sensor; The coordinates of the intersection of the eye tracking line of sight and the first training image are calculated by using a coordinate calculation function preset in the eye tracking sensor to obtain the coordinates of the gaze point where the eye tracking line of sight of the user's dominant eye falls on the first training image.

9. A vision training system based on binocular dichoptic viewing as claimed in claim 6, characterized in that: The image processing module includes a parameter determination unit, an area calculation unit and a Gaussian blur unit; The parameter determination unit is used to determine the convolution kernel value, the two-dimensional direction standard deviation and the radius of the fixation point blur area in the Gaussian blur function according to the corrected vision of the amblyopic eye; The area calculation unit is used to obtain the blurred area of ​​each first training image displayed in the dominant eye display channel based on the blurred area radius of the gaze point and taking the gaze point coordinates as the center; The Gaussian blur unit is used to perform Gaussian blur on the blurred area of ​​each first training image according to the convolution kernel value and the two-dimensional standard deviation through the Gaussian blur function to obtain a second training image after Gaussian blur.

10. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the program is executed by a processor, a visual training method based on binocular dichopia as claimed in any one of claims 1 to 5 is implemented.

Images