Field of view detection method, apparatus and field of view detection device

Abstract

The application relates to a visual field detection method and device and a visual field detection equipment. A plurality of detection data are acquired, focus information of a detection object is obtained according to the plurality of detection data and a preset target visual field detection model, and a visual field detection result of the detection object is obtained according to the focus information; the detection data include visual physiological data of the detection object in a process in which the visual field detection equipment displays a stimulation point, and the focus information includes a concentration result of attention of the detection object and an observation result of the detection object on the stimulation point. In the embodiment of the application, the focus information is obtained based on the visual physiological data of the detection object in the process in which the visual field detection equipment displays the stimulation point, the visual field detection result is generated based on the focus information, that is, the detection object does not need to manually feedback by pressing a button in the case of watching the stimulation point, the visual field detection result is generated according to the feedback result, the complexity of interaction is reduced, subjective error caused by manual feedback is avoided, and the accuracy of the visual field detection result is improved.

Description

Technical Field

[0001] This application relates to the field of medical technology, and in particular to a visual field detection method, device, and equipment. Background Technology

[0002] Visual field testing is a non-invasive examination method that helps detect diseases early and monitor their progression. If the visual field test results show visual field defects, the doctor may conduct further examinations to determine the cause of the visual field defects and select an appropriate treatment plan, such as medication or surgery, based on the type, grade, and stage of the disease.

[0003] In current visual field testing methods, patients need to press the trigger of the visual field testing device when they see the stimulus point displayed on the device to provide feedback. However, patients may have difficulty concentrating or may not be able to correctly perceive the stimulus point, which leads to a high degree of subjectivity in the feedback results and thus lower reliability of the visual field testing results. Summary of the Invention

[0004] Therefore, it is necessary to provide a visual field detection method, apparatus, and device that can improve the accuracy of visual field detection results in response to the above-mentioned technical problems.

[0005] Firstly, this application provides a field-of-view detection method, including:

[0006] Acquire multi-frame detection data; the detection data includes visual physiological data of the detection object during the process of the visual field detection device displaying the stimulus point;

[0007] Based on the multi-frame detection data and the preset target field of view detection model, the focus information of the detected object is obtained; the focus information includes the attention concentration result of the detected object and the observation result of the detected object on the stimulus point;

[0008] The field of view detection result of the detected object is obtained based on the focusing information.

[0009] In one embodiment, obtaining the focus information of the detected object based on the multi-frame detection data and a preset target field-of-view detection model includes:

[0010] The detection data are normalized according to the data type to obtain multiple normalized detection data.

[0011] The focusing information is obtained based on the normalized detection data and the target field of view detection model.

[0012] In one embodiment, the normalized detection data includes the normalized optical axis vectors of the two eyes of the detection object and the normalized fixation point of the visual field detection device; obtaining the focusing information based on the normalized detection data and the target visual field detection model includes:

[0013] For each frame of the normalized optical axis vector of the two eyes, based on the attribute parameters of the visual field detection device and the normalized optical axis vector of the two eyes, the normalized line-of-sight intersection point of the detection object in the visual field detection device is determined.

[0014] The normalized fixed viewpoint offset is determined based on the difference between the normalized line-of-sight intersection and the normalized fixed viewpoint.

[0015] Based on the difference in the normalized optical axis vector of the two eyes of the test subject, the normalized optical axis vector difference of the two eyes of the test subject is determined.

[0016] The normalized detection data, the normalized fixed viewpoint offset, and the normalized optical axis vector difference are input into the target field of view detection model to obtain the focusing information.

[0017] In one embodiment, the step of inputting each of the normalized detection data, each of the normalized fixed viewpoint offsets, and each of the normalized optical axis vector differences into the target field of view detection model to obtain the focusing information includes:

[0018] For each frame of the normalized detection data, the normalized detection data, the normalized fixed viewpoint offset, and the normalized optical axis vector are concatenated to obtain sequence data;

[0019] The position encoding module in the target field detection model is used to encode the order relationship of multiple sequence data to obtain multiple initial feature vectors;

[0020] The feature encoding module in the target field detection model is used to extract features from the multiple initial feature vectors to obtain multiple target feature vectors;

[0021] The focusing information is obtained by linearly mapping multiple target feature vectors.

[0022] In one embodiment, the step of extracting features from multiple initial feature vectors using the feature encoding module in the target vision detection model to obtain each target feature vector includes:

[0023] Multiple initial feature vectors are input into the self-attention submodule of the feature encoding module for correlation processing to obtain multiple embedded feature vectors;

[0024] Multiple embedded feature vectors are input into the feedforward network submodule of the feature encoder module for nonlinear transformation processing to obtain multiple target feature vectors.

[0025] In one embodiment, the method further includes:

[0026] Acquire training data; the training data includes visual physiological data of the test subject during the process of the visual field detection device displaying stimulus points;

[0027] The initial visual field detection model is trained based on the training data and the corresponding standard focusing information to obtain the target visual field detection model. The standard focusing information includes the standard attention concentration result of the test subject and the standard observation result of the test subject on the stimulus point. The standard observation result is determined based on the stimulus intensity of the stimulus point and a preset stimulus intensity threshold. The standard attention concentration result is determined based on the loss of the fixation point, the feedback result of the test subject on the stimulus point, and the standard observation result.

[0028] Secondly, this application also provides a field of view detection device, comprising:

[0029] The acquisition module is used to acquire multi-frame detection data; the detection data includes visual physiological data of the detection object during the process of the visual field detection device displaying stimulus points;

[0030] The first determining module is used to obtain the focusing information of the detected object based on the multi-frame detection data and the preset target field of view detection model; the focusing information includes the attention concentration result of the detected object and the observation result of the detected object on the stimulus point;

[0031] The second determining module is used to obtain the field of view detection result of the detection object based on the focusing information.

[0032] Thirdly, this application also provides a field of view detection device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0033] Acquire multi-frame detection data; the detection data includes visual physiological data of the detection object during the process of the visual field detection device displaying the stimulus point;

[0034] Based on the multi-frame detection data and the preset target field of view detection model, the focus information of the detected object is obtained; the focus information includes the attention concentration result of the detected object and the observation result of the detected object on the stimulus point;

[0035] The field of view detection result of the detected object is obtained based on the focusing information.

[0036] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0037] Acquire multi-frame detection data; the detection data includes visual physiological data of the detection object during the process of the visual field detection device displaying the stimulus point;

[0038] Based on the multi-frame detection data and the preset target field of view detection model, the focus information of the detected object is obtained; the focus information includes the attention concentration result of the detected object and the observation result of the detected object on the stimulus point;

[0039] The field of view detection result of the detected object is obtained based on the focusing information.

[0040] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0041] Acquire multi-frame detection data; the detection data includes visual physiological data of the detection object during the process of the visual field detection device displaying the stimulus point;

[0042] Based on the multi-frame detection data and the preset target field of view detection model, the focus information of the detected object is obtained; the focus information includes the attention concentration result of the detected object and the observation result of the detected object on the stimulus point;

[0043] The field of view detection result of the detected object is obtained based on the focusing information.

[0044] The aforementioned visual field detection method, apparatus, and device acquire multiple frames of detection data, obtain the focusing information of the detected object based on the multiple frames of detection data and a preset target visual field detection model, and obtain the visual field detection result of the detected object based on the focusing information. The detection data includes the visual physiological data of the detected object during the process of the visual field detection device displaying the stimulus point, and the focusing information includes the detection object's attention concentration result and the detection object's observation result of the stimulus point. In the embodiments of this application, the attention concentration result and observation result are obtained based on the visual physiological data of the detected object during the process of the visual field detection device displaying the stimulus point and the target visual field detection model. The visual field detection result is generated based on the attention concentration result and observation result. That is, there is no need for the detected object to press a button for manual feedback when viewing the stimulus point. The visual field detection result is generated based on the feedback result, which greatly reduces the complexity of the interaction, avoids the subjective error caused by manual feedback, and improves the accuracy of the visual field detection result. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 This is an application environment diagram of the field-of-view detection method in one embodiment;

[0047] Figure 2 This is a flowchart illustrating a field-of-view detection method in one embodiment;

[0048] Figure 3 This is a flowchart illustrating the field-of-view detection method in another embodiment;

[0049] Figure 4 This is a flowchart illustrating a method for determining focus information in one embodiment;

[0050] Figure 5 This is a flowchart illustrating the focus information determination method in another embodiment;

[0051] Figure 6 This is a flowchart illustrating the focus information determination method in another embodiment;

[0052] Figure 7 This is a schematic diagram of a target field-of-view detection model in one embodiment;

[0053] Figure 8 This is a flowchart illustrating a method for determining target feature vectors in one embodiment;

[0054] Figure 9 This is a flowchart illustrating a method for determining a target field of view detection model in one embodiment;

[0055] Figure 10 This is a flowchart illustrating the field-of-view detection method in another embodiment;

[0056] Figure 11 This is a structural block diagram of a field-of-view detection device in one embodiment. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0058] The field-of-view detection method provided in this application embodiment can be applied to, for example... Figure 1The application environment shown includes a field-of-view detection device, the internal structure of which can be illustrated as follows. Figure 1 As shown, the field of view detection device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computational and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, or other technologies. When the computer program is executed by the processor, it implements a field of view detection method. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be an LCD screen, and the input device of the field of view detection device can be an external touch screen, or a button, trackball, or touchpad set on the housing of the field of view detection device, or an external keyboard, touchpad, or mouse, etc.

[0059] Optionally, the field of view detection device can be a wearable type, such as integrating field of view detection into a head-mounted device, which can be a virtual reality (VR) device, an augmented reality (AR) device, a mixed reality (MR) device, or smart glasses. The field of view detection device can also be a non-wearable type, such as a desktop detection device.

[0060] Those skilled in the art will understand that Figure 1 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the field of view detection device to which the present application is applied. A specific field of view detection device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0061] In one exemplary embodiment, such as Figure 2 As shown, a field-of-view detection method is provided, which is applied to... Figure 1 The following explanation uses a field-of-view detection device as an example, including the following steps S201 to S203. Wherein:

[0062] S201, acquire multi-frame detection data; the detection data includes visual physiological data of the detection object during the process of the visual field detection device displaying the stimulus point.

[0063] The detection data includes visual physiological data of the test subject during the display of stimulus points on the visual field detection device. This visual physiological data reflects the test subject's cognitive state and attention level, and is not solely limited to visual stimuli. For example, pupil size is related to the level of psychological load and the intensity of light stimulation; blinking can block visual input to reduce interference, and internal creative thinking often leads to a higher blink frequency and longer blink duration; convergence refers to the angle between the right and left eye's focusing lines, and it can be used to determine the change in the test subject's depth of focus on visual input. Therefore, acquiring the visual physiological data of the test subject across multiple frames before and after the display of the nth stimulus point could be based on the pupil diameter r. t Detection of the optical axis vectors of the subject's eyes , Blinking frequency b t and palpebral fissure height o t And so on. t represents the time corresponding to the detection data, which is also the number of frames corresponding to the detection data.

[0064] Optionally, the detection data may also include basic data of the detection subject. For example, the subject's age 'a', the subject's interpupillary distance information 'p', etc. d The detection data may also include data related to the field of view (GLA) detection equipment, such as the coordinates of the fixed viewpoint of the GLA detection equipment. The time t for the nth stimulus point to appear n The relative position of the nth stimulus point The stimulus intensity L at the nth stimulus point n .

[0065] Optionally, the detection data may also include electroencephalogram (EEG) data of the subject during the display of stimulus points on the visual field testing device.

[0066] In this embodiment, the detection data can be sent to the visual field detection device by other detection devices, obtained from the corresponding system in the hospital, collected by the visual field detection device, or input by the user. For example, the basic data of the detection object can be input by the user or imported from the hospital system; the visual physiological data can be obtained by the wearable visual field detection device directly collecting data from the detection object.

[0067] Combination Figure 3As shown, taking a head-mounted visual field testing device as an example, the device is started to ensure that all systems function normally. Personalized parameter adjustments are made based on the subject's age and interpupillary distance (IPD) information. The post-processing platform of the head-mounted visual field testing device determines the fixation point based on the IPD information and the stimulus point based on the age information, and outputs the fixation point and stimulus point on the display screen. Simultaneously, the head-mounted visual field testing device collects the subject's visual physiological data.

[0068] S202, based on multi-frame detection data and a preset target field of view detection model, obtain the focus information of the detected object; the focus information includes the attention concentration result of the detected object and the observation result of the detected object on the stimulus point.

[0069] Optionally, the target vision detection model can be an autoencoder, a long short-term memory network, or a convolutional neural network, etc.

[0070] In this embodiment, the above is continued. Figure 3 As shown, multi-frame detection data can be directly input into the target visual field detection model to obtain the attention concentration result of the detected object and the observation result of the detected object on the stimulus point. The attention concentration result is the degree of attention concentration of the detected object, and the observation result is the probability that the detected object sees the stimulus point.

[0071] In one possible implementation, multi-frame detection data can be preprocessed, and the preprocessed data can be input into the target field-of-view detection model to obtain focus information. For example, the detection data can be normalized, outlier data can be removed, or duplicate data can be removed.

[0072] In one possible implementation, since the stimulation point includes multiple stimulation points that appear at different times, the relative position and stimulation intensity of the stimulation point at the next moment can be adjusted in real time based on the focusing information, and the focusing information can be determined subsequently based on the stimulation point after the last adjustment and other detection data.

[0073] S203, obtain the field of view detection result of the detected object based on the focusing information.

[0074] In this embodiment, when the focusing information meets preset requirements, the visual field detection result of the detection object is generated based on the relative position and intensity of the stimulus point. For example, the focusing information includes the detection object's attention concentration result and the detection object's observation result of the stimulus point. The preset requirements can be an attention threshold of 60% and a probability threshold of 70% for the detection object to see the stimulus point. For a given stimulus point, if the detection object's attention concentration result is greater than 60% and the observation result is greater than 70%, the relative position and intensity of the stimulus point are saved. Based on the saved stimulus intensity and relative position of the stimulus point, the visual field detection result of the detection object is generated.

[0075] In the aforementioned visual field detection method, multiple frames of detection data are acquired. Based on the multiple frames of detection data and a preset target visual field detection model, the focusing information of the detection object is obtained. The visual field detection result of the detection object is then obtained based on the focusing information. The detection data includes the visual physiological data of the detection object during the display of the stimulus point on the visual field detection device. The focusing information includes the detection object's attention concentration result and the detection object's observation result of the stimulus point. In this embodiment, the attention concentration result and observation result are obtained based on the visual physiological data of the detection object during the display of the stimulus point on the visual field detection device and the target visual field detection model. The visual field detection result is generated based on the attention concentration result and observation result. That is, the detection object does not need to press a button for manual feedback when viewing the stimulus point. The visual field detection result is generated based on the feedback result, greatly reducing the complexity of the interaction, avoiding the subjective error caused by manual feedback, and improving the accuracy of the visual field detection result.

[0076] Figure 4 This is a flowchart illustrating a method for determining focus information in one embodiment, such as... Figure 4 As shown, this application embodiment relates to a possible implementation of how to obtain the focus information of the detected object based on multi-frame detection data and a preset target field-of-view detection model, including the following steps:

[0077] S401, normalize each detection data according to its data type to obtain multiple normalized detection data.

[0078] In this embodiment, each detection data is normalized according to its data type to obtain multiple normalized detection data. For example, if the detection data is age, it can be normalized by dividing the age by 150 to obtain the normalized age; if the detection data is pupil diameter, the pupil diameter of the test subject is normalized based on the maximum pupil diameter when the stimulus point is displayed on the visual field testing device to obtain the normalized pupil diameter. The detection data is blink frequency. The blink frequency of each test subject is normalized based on the maximum blink frequency when the stimulus point is displayed on the visual field detection device, resulting in the normalized blink frequency. .

[0079] S402 obtains focusing information based on multiple normalized detection data and target field-of-view detection models.

[0080] In this embodiment, multiple normalized detection data can be input into the target visual field detection model to obtain focusing information. Alternatively, new data can be generated based on the multiple normalized detection data, and then input into the target visual field detection model along with the new data to obtain focusing information. For example, the normalized detection data includes the normalized optical axis vectors of the two eyes of the detection object. The normalized optical axis vector difference can be obtained based on the normalized optical axis vectors of the two eyes, and the normalized detection data and the normalized optical axis vector difference can be input into the target visual field detection model to obtain focusing information.

[0081] In this embodiment, each detection data is normalized according to its data type to obtain multiple normalized detection data. Based on the multiple normalized detection data and the target field of view detection model, focusing information is obtained, which can improve the efficiency and accuracy of focusing information determination.

[0082] Figure 5 This is a flowchart illustrating the focus information determination method in another embodiment, such as... Figure 5 As shown, this application embodiment relates to a possible implementation of how to obtain focus information based on normalized detection data and a target field of view detection model, including the following steps:

[0083] S501, for each frame of the normalized optical axis vector of both eyes, based on the attribute parameters of the visual field detection device and the normalized optical axis vector of both eyes, determine the normalized line-of-sight intersection point of the detected object on the visual field detection device.

[0084] In this embodiment, the attribute parameters of the field-of-view detection device include the simulated lens parameters and the display screen parameters. For example, the normalized optical axis vector at time t is used as follows: , Based on the simulated lens parameters of the field of view detection device, the intersection points of the lines of sight of the object being detected on the field of view detection device are calculated. and .

[0085] The line-of-sight intersection point is normalized based on the parameters displayed on the visual field detection device. The normalized line-of-sight intersection point is: , w and h represent the width and height of the display screen of the field of view detection device in pixels.

[0086] S502, determine the normalized fixed viewpoint offset based on the difference between the normalized line-of-sight intersection and the normalized fixed viewpoint.

[0087] In this embodiment, the difference between the normalized line-of-sight intersection point and the normalized fixed viewpoint can be used to obtain the normalized fixed viewpoint offset, i.e. , .

[0088] S503, Based on the difference in the normalized optical axis vector of the two eyes of the test subject, determine the normalized optical axis vector difference of the two eyes of the test subject.

[0089] In this embodiment, the normalized optical axis vectors of the two eyes of the detection object are... , The difference is used as the normalized optical axis vector difference between the two eyes of the test subject. That is... .

[0090] S504 inputs each normalized detection data, each normalized fixed viewpoint offset, and each normalized optical axis vector difference into the target field of view detection model to obtain focusing information.

[0091] In this embodiment, the normalized detection data, normalized fixed viewpoint offset, and normalized optical axis vector difference corresponding to each frame are concatenated to obtain sequence data. Multiple sequence data are input into the target field of view detection model to obtain focusing information.

[0092] In this embodiment, for each frame of normalized optical axis vectors of both eyes, based on the attribute parameters of the visual field detection device and the normalized optical axis vectors of both eyes, the normalized line-of-sight intersection point of the detection object on the visual field detection device is determined. Based on the difference between the normalized line-of-sight intersection point and the normalized fixed gaze point, the normalized fixed gaze point offset is determined. Based on the difference between the normalized optical axis vectors of the detection object's two eyes, the normalized optical axis vector difference of the detection object's two eyes is determined. Each normalized detection data point, each normalized fixed gaze point offset, and each normalized optical axis vector difference are input into the target visual field detection model to obtain focusing information. This embodiment generates the normalized optical axis vector difference and the normalized fixed gaze point offset based on the normalized fixed gaze point and normalized optical axis vector, expanding the normalized detection data, enriching the dataset, and thus improving the accuracy of focusing information determination.

[0093] Figure 6 This is a flowchart illustrating the focus information determination method in another embodiment, such as... Figure 6 As shown, this application embodiment relates to a possible implementation of how to input each normalized detection data, each normalized fixed viewpoint offset, and each normalized optical axis vector difference into a target field of view detection model to obtain focusing information, including the following steps:

[0094] S601, for each frame of normalized detection data, the normalized detection data, the normalized fixed viewpoint offset, and the normalized optical axis vector difference are concatenated to obtain sequence data.

[0095] In this embodiment, normalized detection data, normalized fixation point offset, and normalized optical axis vector difference are concatenated to obtain sequence data. For example, the normalized data includes normalized age, normalized interpupillary distance, normalized pupil diameter, normalized blink frequency, normalized palpebral fissure height, normalized optical axis vectors of both eyes, normalized fixation point, normalized relative position of stimulus points, normalized stimulus intensity of stimulus points, and stimulus point occurrence time. Concatenating the normalized detection data, normalized fixation point offset, and normalized optical axis vector difference yields sequence data with a dimension of 16.

[0096] S602 uses the position encoding module in the target vision detection model to encode the order relationship of multiple sequence data to obtain multiple initial feature vectors.

[0097] In this embodiment, Figure 7 This is a schematic diagram of a target field-of-view detection model in one embodiment, such as... Figure 7 As shown, since the temporal information of sequence data is important, the position encoding module is used to encode the order relationship of multiple sequence data, that is, to label multiple sequence data and use the labeled sequence data as the initial feature vector.

[0098] S603 uses the feature encoding module in the target vision detection model to extract features from multiple initial feature vectors, thereby obtaining multiple target feature vectors.

[0099] In this embodiment, the feature encoding module may include a self-attention submodule and a feedforward network submodule. The self-attention submodule and the feedforward network submodule are used to extract features from multiple initial feature vectors to obtain multiple target feature vectors.

[0100] In one possible implementation, the feature encoding module may also include convolutional layers and pooling layers, which are used to extract features from multiple initial feature vectors to obtain the target feature vector.

[0101] Optionally, the feature encoding module may include multiple modules.

[0102] S604 performs a linear mapping on the target feature vector to obtain focusing information.

[0103] In this embodiment, since the focusing information includes attention concentration results and observation results, a linear mapping layer can be used to linearly map the target feature vector into a two-dimensional vector with a real value range, and then a sigmoid layer can be used to map the two-dimensional vector to [0, 1). The output of the sigmoid layer... This refers to the probability of the subject focusing their attention and the probability of seeing the stimulus point.

[0104] In this embodiment, for each frame of normalized detection data, the normalized detection data, normalized fixed viewpoint offset, and normalized optical axis vector difference are concatenated to obtain sequence data. The position encoding module in the target field of view detection model encodes the sequential relationship of multiple sequence data to obtain multiple initial feature vectors. The feature encoding module in the target field of view detection model extracts features from these initial feature vectors to obtain target feature vectors. Linear mapping is then performed on these target feature vectors to obtain focus information. This embodiment utilizes the position encoding module to encode the sequential relationship of multiple sequence data, fully considering the temporal information of multiple sequence data, thus improving the accuracy of focus information determination. Furthermore, the use of the feature encoding module to obtain feature vectors fully considers the practicality of the target field of view detection model.

[0105] Figure 8 This is a flowchart illustrating a target feature vector determination method in one embodiment, as shown below. Figure 8 As shown, this application embodiment relates to a possible implementation of how to use the feature encoding module in a target vision detection model to extract features from multiple initial feature vectors to obtain multiple target feature vectors, including the following steps:

[0106] S801 inputs multiple initial feature vectors into the self-attention submodule in the feature encoding module for correlation processing to obtain multiple embedded feature vectors.

[0107] In this embodiment, since there are correlations between data of different dimensions within the same sequence data, there are also correlations between data of the same dimension in different sequences data. For example, the pupil size of the detected object is strongly correlated with the location of the stimulus point and the stimulus intensity.

[0108] By utilizing the self-attention mechanism of the self-attention submodule, we can find the correlation between different initial feature vectors and the correlation within the same initial feature vector itself, thus obtaining multiple embedded feature vectors.

[0109] S802 inputs multiple embedded feature vectors into the feedforward network submodule in the feature encoder module for nonlinear transformation processing to obtain multiple target feature vectors.

[0110] In this embodiment, multiple embedded feature vectors are input to the feedforward network submodule. The feedforward network submodule performs a nonlinear transformation on the multiple embedded feature vectors to further extract the features of the embedded feature vectors and obtain multiple target feature vectors.

[0111] If there are multiple feature encoding modules, after extracting features by performing nonlinear transformations on multiple embedded feature vectors, the vectors are input into the next feature encoding module, and the feature vector output by the last feature encoding module is used as the target feature vector.

[0112] In this embodiment, multiple initial feature vectors are input into the self-attention submodule of the feature encoding module for correlation processing to obtain multiple embedded feature vectors. These multiple embedded feature vectors are then input into the feedforward network submodule of the feature encoder module for nonlinear transformation processing to obtain multiple target feature vectors. This fully considers the correlation between the sequence data and improves the accuracy of target feature vector acquisition, thereby improving the accuracy of focusing information obtained based on the target feature vectors.

[0113] In one embodiment, a method for determining a target visual field detection model is also provided, comprising: acquiring training data; the training data including visual physiological data of a test subject during the display of a stimulus point on a visual field detection device; training an initial visual field detection model based on the training data and the standard focusing information corresponding to the training data to obtain a target visual field detection model; the standard focusing information including the test subject's standard attention concentration result and the test subject's standard observation result of the stimulus point; the standard observation result being determined based on the stimulus intensity of the stimulus point and a preset stimulus intensity threshold; and the standard attention concentration result being determined based on the loss of fixation point, the test subject's feedback result of the stimulus point, and the standard observation result.

[0114] The standard focus information includes the test subject's standard attention concentration results and the test subject's standard observation results of the stimulus point. The standard observation results are based on the stimulus intensity of the stimulus point. and preset stimulus intensity threshold Whether the test subject sees the stimulus point is determined based on the stimulus intensity of the stimulus point and a preset stimulus intensity threshold. The preset stimulus intensity threshold is determined based on the stimulus intensity at the corresponding position in each frame of training data of the test subject.

[0115] like Figure 9 As shown, the standard attention concentration results are based on the loss of fixation point. Feedback results of the test subjects to the stimulation points Determined by the standard observation result s, The standard attention result is 1 if the test subject sees the stimulus and reports seeing it, and the fixation point is not lost; or, if the test subject does not see the stimulus and does not report seeing it, and the fixation point is not lost, the standard attention result is 1; if the test subject sees the stimulus but does not report seeing it, the standard attention result is 0; if the test subject does not see the stimulus but reports seeing it, the standard attention result is 0; if the fixation point is lost, the standard attention result is 0.

[0116] In this embodiment, continue as follows Figure 9 As shown, the training data includes basic data of the test subjects, visual physiological data, fixation point data, and stimulus point data, as well as the optical axis vector difference and fixation point offset obtained based on the binocular optical axis vectors in the fixation point and visual physiological data. The training data can be normalized or unnormalized. The training data is input into the initial visual field detection model to obtain predicted focus information. The initial visual field detection model is then trained based on the loss values ​​of the predicted focus information and the standard focus information, and the target visual field detection model is obtained by minimizing the loss value.

[0117] Optionally, the binary cross-entropy loss function can be used to calculate the loss value L between the predicted focus information and the standard focus information. Where q is the number of training data in one iteration, y is the standard attention concentration result or standard observation result in the training data, and p y To predict the outcome of attention concentration or to predict the results of observation.

[0118] In this embodiment, training data is acquired, and an initial field-of-view detection model is trained based on the training data and the standard focusing information corresponding to the training data to obtain a target field-of-view detection model, which lays the foundation for obtaining the focusing information of the detected object based on the target field-of-view detection model.

[0119] Figure 10 This is a flowchart illustrating the field-of-view detection method in another embodiment, as shown below. Figure 10 As shown, it includes the following steps:

[0120] S1001, acquire multi-frame detection data; the detection data includes visual physiological data of the detected object during the process of the visual field detection device displaying the stimulus point;

[0121] S1002, Normalize each detection data according to the data type to obtain multiple normalized detection data; The normalized detection data includes the normalized optical axis vectors of the two eyes of the detection object and the normalized fixation point of the visual field detection device;

[0122] S1003, For each frame of normalized optical axis vectors of both eyes, based on the attribute parameters of the visual field detection device and the normalized optical axis vectors of both eyes, determine the normalized line-of-sight intersection point of the detection object on the visual field detection device.

[0123] S1004, Determine the normalized fixed viewpoint offset based on the difference between the normalized line-of-sight intersection and the normalized fixed viewpoint;

[0124] S1005, Based on the difference in the normalized optical axis vector of the two eyes of the test subject, determine the normalized optical axis vector difference of the two eyes of the test subject.

[0125] S1006, For each frame of normalized detection data, the normalized detection data, the normalized fixed viewpoint offset, and the normalized optical axis vector difference are concatenated to obtain sequence data;

[0126] S1007, the position encoding module in the target vision detection model is used to encode the order relationship of multiple sequence data to obtain multiple initial feature vectors;

[0127] S1008, multiple initial feature vectors are input into the self-attention submodule in the feature encoding module for correlation processing to obtain multiple embedded feature vectors;

[0128] S1009, multiple embedded feature vectors are input into the feedforward network submodule in the feature encoder module for nonlinear transformation processing to obtain multiple target feature vectors;

[0129] S1010 performs linear mapping on multiple target feature vectors to obtain focusing information;

[0130] S1011, obtain the field of view detection result of the detected object based on the focusing information.

[0131] In this embodiment, attention concentration results and observation results are obtained based on the visual physiological data of the detection object during the display of the stimulus point on the visual field detection device and the target visual field detection model. Visual field detection results are generated based on the attention concentration results and observation results. That is, the detection object does not need to press a button to manually provide feedback when viewing the stimulus point. Visual field detection results are generated based on the feedback results, which greatly reduces the complexity of the interaction, avoids the subjective error caused by manual feedback, and improves the accuracy of visual field detection results.

[0132] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0133] Based on the same inventive concept, this application also provides a field of view detection device for implementing the field of view detection method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more field of view detection device embodiments provided below can be found in the limitations of the field of view detection method described above, and will not be repeated here.

[0134] In one exemplary embodiment, such as Figure 11 As shown, a field of view detection device is provided, including: an acquisition module 11, a first determination module 12, and a second determination module 13, wherein:

[0135] The acquisition module 11 is used to acquire multi-frame detection data; the detection data includes visual physiological data of the detection object during the process of the visual field detection device displaying the stimulus point;

[0136] The first determining module 12 is used to obtain the focusing information of the detected object based on multi-frame detection data and a preset target field of view detection model; the focusing information includes the attention concentration result of the detected object and the observation result of the detected object on the stimulus point;

[0137] The second determining module 13 is used to obtain the field of view detection result of the detected object based on the focusing information.

[0138] In one embodiment, the first determining module 12 is specifically used to normalize each detection data according to the data type to obtain multiple normalized detection data; and to obtain focusing information based on the multiple normalized detection data and the target field of view detection model.

[0139] In one embodiment, the first determining module 12 is specifically used to determine, for each frame of normalized optical axis vectors of both eyes, the normalized line-of-sight intersection point of the detected object on the visual field detection device based on the attribute parameters of the visual field detection device and the normalized optical axis vectors of both eyes; determine the normalized fixed-viewpoint offset based on the difference between the normalized line-of-sight intersection point and the normalized fixed-viewpoint; determine the normalized optical axis vector difference between the two eyes of the detected object based on the difference between the normalized optical axis vectors of the two eyes of the detected object; and input each normalized detection data, each normalized fixed-viewpoint offset, and each normalized optical axis vector difference into the target visual field detection model to obtain focusing information.

[0140] In one embodiment, the first determining module 12 is specifically used to, for each frame of normalized detection data, concatenate the normalized detection data, the normalized fixed viewpoint offset, and the normalized optical axis vector difference to obtain sequence data; encode the order relationship of multiple sequence data using the position encoding module in the target field of view detection model to obtain multiple initial feature vectors; extract features from the multiple initial feature vectors using the feature encoding module in the target field of view detection model to obtain target feature vectors; and perform linear mapping on the target feature vectors to obtain focusing information.

[0141] In one embodiment, the first determining module 12 is specifically used to input each initial feature vector into the self-attention submodule in the feature encoding module for correlation processing to obtain an embedded feature vector; and to input the embedded feature vector into the feedforward network submodule in the feature encoder module for nonlinear transformation processing to obtain a target feature vector.

[0142] In one embodiment, the field-of-view detection device further includes:

[0143] The data acquisition module is used to acquire training data; the training data includes the visual physiological data of the test subjects during the process of the visual field detection device displaying stimulus points;

[0144] The training module is used to train the initial visual field detection model based on the training data and the corresponding standard focusing information to obtain the target visual field detection model. The standard focusing information includes the standard attention concentration results of the test subject and the standard observation results of the test subject on the stimulus point. The standard observation results are determined based on the stimulus intensity of the stimulus point and the preset stimulus intensity threshold. The standard attention concentration results are determined based on the loss of fixation point, the feedback results of the test subject on the stimulus point, and the standard observation results.

[0145] Each module in the aforementioned field of view detection device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of the field of view detection device in hardware form or independent of it, or they can be stored in the memory of the field of view detection device in software form, so that the processor can call and execute the operations corresponding to each module.

[0146] In one exemplary embodiment, a field-of-view detection device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of any of the above method embodiments.

[0147] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the above method embodiments.

[0148] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of any of the above method embodiments.

[0149] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0150] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0151] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0152] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for detecting a field of view, characterized in that, The method includes: Acquire multi-frame detection data; the detection data includes visual physiological data of the detection object during the process of the visual field detection device displaying stimulus points; the visual physiological data is used to characterize the cognitive state and attention level of the detection object; Based on the multi-frame detection data and the preset target visual field detection model, the focus information of the detected object is obtained; the focus information includes the attention concentration result of the detected object and the observation result of the detected object on the stimulus point; the observation result is the probability that the detected object observes the stimulus point; When the focusing information meets the preset requirements, the visual field detection result of the detection object is generated based on the relative position of the stimulation point and the stimulation intensity.

2. The method according to claim 1, characterized in that, The step of obtaining the focus information of the detected object based on the multi-frame detection data and the preset target field-of-view detection model includes: The detection data are normalized according to the data type to obtain multiple normalized detection data. The focusing information is obtained based on the normalized detection data and the target field of view detection model.

3. The method according to claim 2, characterized in that, The normalized detection data includes the normalized optical axis vectors of the two eyes of the detection object and the normalized fixed point of the visual field detection device; The step of obtaining the focusing information based on the normalized detection data and the target field-of-view detection model includes: For each frame of the normalized optical axis vector of the two eyes, based on the attribute parameters of the visual field detection device and the normalized optical axis vector of the two eyes, the normalized line-of-sight intersection point of the detected object on the visual field detection device is determined. The normalized fixed viewpoint offset is determined based on the difference between the normalized line-of-sight intersection and the normalized fixed viewpoint. Based on the difference in the normalized optical axis vector of the two eyes of the test subject, the normalized optical axis vector difference of the two eyes of the test subject is determined. The normalized detection data, the normalized fixed viewpoint offset, and the normalized optical axis vector difference are input into the target field of view detection model to obtain the focusing information.

4. The method according to claim 3, characterized in that, The step of inputting each of the normalized detection data, each of the normalized fixed viewpoint offsets, and each of the normalized optical axis vector differences into the target field of view detection model to obtain the focusing information includes: For each frame of the normalized detection data, the normalized detection data, the normalized fixed viewpoint offset, and the normalized optical axis vector difference are concatenated to obtain sequence data; The position encoding module in the target field detection model is used to encode the order relationship of multiple sequence data to obtain multiple initial feature vectors; The feature encoding module in the target field detection model is used to extract features from the multiple initial feature vectors to obtain multiple target feature vectors; The focusing information is obtained by linearly mapping multiple target feature vectors.

5. The method according to claim 4, characterized in that, The step of extracting features from multiple initial feature vectors using the feature encoding module in the target field detection model to obtain each target feature vector includes: Multiple initial feature vectors are input into the self-attention submodule of the feature encoding module for correlation processing to obtain multiple embedded feature vectors; Multiple embedded feature vectors are input into the feedforward network submodule of the feature encoder module for nonlinear transformation processing to obtain multiple target feature vectors.

6. The method according to claim 3, characterized in that, The method further includes: Acquire training data; the training data includes visual physiological data of the test subject during the process of the visual field detection device displaying stimulus points; The initial visual field detection model is trained based on the training data and the corresponding standard focusing information to obtain the target visual field detection model. The standard focusing information includes the standard attention concentration result of the test subject and the standard observation result of the test subject on the stimulus point. The standard observation result is determined based on the stimulus intensity of the stimulus point and a preset stimulus intensity threshold. The standard attention concentration result is determined based on the loss of the fixation point, the feedback result of the test subject on the stimulus point, and the standard observation result.

7. A field of view detection device, characterized in that, The device includes: The acquisition module is used to acquire multi-frame detection data; the detection data includes visual physiological data of the detection object during the process of the visual field detection device displaying stimulus points; the visual physiological data is used to characterize the cognitive state and attention level of the detection object; The first determining module is used to obtain the focusing information of the detected object based on the multi-frame detection data and a preset target visual field detection model; the focusing information includes the attention concentration result of the detected object and the observation result of the detected object on the stimulus point; the observation result is the probability that the detected object observes the stimulus point; The second determining module is used to generate the visual field detection result of the detection object based on the relative position of the stimulation point and the stimulation intensity when the focusing information meets the preset requirements.

8. A field-of-view detection device, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A wearable field-of-view detection device, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

10. The device according to claim 9, characterized in that, The device includes any one of virtual reality devices, augmented reality devices, mixed reality devices, and smart glasses.

Images