Regional human eye sight line calibration method for people with defect of visual field

Abstract

The invention provides a regional human eye sight line calibration method for a visual field defect crowd. The method is used for accurate sight line drop point estimation when the visual field defect crowd uses a head-mounted or telemetering type sight line tracking device. The method comprises the steps of initial calibration, view region division, regional calibration and the like. Initial calibration is carried out, view measurement is carried out on the crowd with the view defect through a view meter and a grid, an approximate visible area and an approximate invisible area are obtained, and mapping calibration is carried out on the normal view. Dividing view areas, obtaining a fixation point distribution diagram, and scanning and combining the view areas into a normal view, a distorted view and an invisible view according to the fixation point distribution condition. And carrying out regional calibration, defining an invisible view as a view boundary, and respectively calibrating a normal view and a distorted view. According to the method, the sight line calibration problem of the visual field defect crowd in different application scenes is solved, the full-region division of the real visual field of the visual field defect crowd is realized, and the method has relatively high accuracy and relatively good robustness.

Description

Technical Field

[0001] The present invention relates to the technical field of human-computer interaction, and in particular to a method for calibrating the sight lines of different regions of a person with visual field impairment. Background Art

[0002] People with visual field loss have lost some of their visual abilities due to congenital or acquired diseases, which often manifests as blind spots or distorted areas in their visual field. Real objects in reality cannot be correctly described in the eyes of people with visual field loss, and their normal gaze habits are also changed. Currently, eye trackers and other equipment are usually unable to perform specific calibration for people with visual field loss, so it is still difficult to study the visual behavior of people with visual field loss through gaze points.

[0003] Currently, the common sight line calibration methods on the market are all aimed at people with normal vision. The gaze point calibration of people with normal vision often only needs to complete the uniform mapping of the eye coordinate system to the screen coordinate system, and does not take into account the distortion of the gaze point or the special behavior of the gaze point that cannot be observed. This sight line calibration method is of no use to people with visual field defects, and research on specific populations will also have problems such as complex operation and insufficient accuracy.

[0004] Therefore, taking into account the particularity of the visual field of people with visual field impairment, it is possible to classify the visual field areas and perform gaze point calibration with different methods for different types of areas. In this way, while achieving high-quality line of sight calibration, it is also possible to integrate the special circumstances of individuals with visual field impairment, which is convenient for behavioral research and diagnosis and treatment of individuals. Summary of the invention

[0005] According to the technical problems mentioned in the above background technology, a method for calibrating the human eye sight line in different regions for people with visual field impairment is provided. The present invention mainly provides a method for calibrating the human eye sight line in different regions for people with visual field impairment. The present invention is mainly used for accurately estimating the sight point when people with visual field impairment use head-mounted or telemetry sight tracking devices. It can complete high-quality sight line calibration for people with different types of visual field impairment, understand the visual behavior and attention allocation of people with visual field impairment, and has positive significance for research in other fields of people with visual field impairment.

[0006] The technical means adopted by the present invention are as follows:

[0007] A method for calibrating the sight line of human eyes in different regions for people with visual field impairment, characterized by comprising the following steps:

[0008] Step 1: Initial calibration: measuring the visual field of people with visual field loss by using a perimeter and a grid to obtain a visible area I and an invisible area I, and mapping and calibrating the normal visual field in the visible area I; the visible area I includes a visual field distortion area and a normal area;

[0009] Step 2: Divide the visual field area; obtain the gaze point distribution map according to the predetermined acquisition method, and scan and merge the visual field area into normal visual field, distorted visual field and invisible visual field according to the distribution of gaze points;

[0010] Step 3: Calibrate by region; define the invisible field of view as the field of view boundary and do not process it; calibrate the normal field of view and the distorted field of view respectively; the normal field of view is calibrated by plane mapping; the distorted field of view is triangulated, and the affine transformation parameters of the local triangulation area are solved, and the triangular areas with roughly the same transformation matrix are merged, and the affine transformation parameters of the merged area are solved;

[0011] Step 4: After calibration is completed, the gaze point and calibration point of the distorted area are compared again. If the difference is within the threshold range, the calibration is considered complete. If the difference is greater than the threshold range, triangulation is performed again, and the centers of the three newly generated triangles are used as new calibration points until the difference is within the threshold range.

[0012] Furthermore, in the initial calibration process, firstly, the largest normal visual field area is framed in the visual field result, and calibration points are set at the four vertices and the center point of the framed area respectively;

[0013] The user is required to keep their eyes open and stare at the calibration points when they appear. The camera is used to collect eye information, and the eye information data is integrated into the 3D eye model to calculate the gaze data before calibration. The parameters of the gaze tracking method are fine-tuned by comparing multiple sets of coordinate data of the gaze points with those of the calibration points.

[0014] Furthermore, during the calibration process of the initial calibration, it is necessary to measure the features of the user's eyes and calculate the sight line data together with the features of the user's eyes and the 3D eye model.

[0015] Furthermore, the field of view area division process includes the following steps:

[0016] Step 21: The user sequentially gazes at evenly arranged gaze points generated by the program;

[0017] Step 22: The user reports the location of the current fixation point according to the grid information. If no fixation point is observed, the record is empty. The fixation point information of the field-deficient population is obtained by observing all fixation points.

[0018] Step 23: Divide the gaze point information of the field-deficient population into normal visual field, distorted visual field and invisible visual field according to the arrangement rules of the gaze points.

[0019] Furthermore, the normal field of vision appears as an orderly and uniform arrangement of gaze points in the gaze point distribution map; the distorted field of vision appears as a chaotic arrangement of gaze points in the gaze point distribution map; and the invisible field of vision appears as a missing gaze point in the gaze point distribution map.

[0020] Furthermore, in step 3, for the normal field of view correction process, the normal field of view mapping is achieved through a polynomial; let P m =[x,y,z] T ,Right now:

[0021]

[0022] in, β=[β x β y β z ] represents mapping parameters; P m Indicates the coordinate position; x, y, z represent the coordinates of the x, y, and z axes respectively; calibration points are generated at the four vertices of the normal field of view, and the user is required to look at the calibration points normally. After the camera captures the eye image and obtains the offset data, the program can be adjusted to compensate for the offset, thereby completing the line of sight calibration of the normal field of view.

[0023] Furthermore, in the step 3, for the calibration process of the distorted field of view, a linear transformation is performed on a vector space in geometry by an affine transformation and a translation is added, thereby transforming it into another vector space;

[0024] Assume that in the original vector space k = (x, y), j = (x ′ ,y ′ ), the conversion between the two vector spaces is realized through the parameter matrix M, the formula is:

[0025] [x ′ ,y ′ ]=M[x,y];

[0026] If the parameter matrices obtained for adjacent triangular regions are roughly the same, it means that the transformation rules of the two triangular regions are consistent, and the two triangular regions are merged.

[0027] Furthermore, the threshold range is 1 percent of the size of the long side of the screen.

[0028] Furthermore, the fine-tuning refers to automatically adjusting the coordinate parameters generated by the model when calculating the overall center coordinates of the densely distributed area of ​​the gaze points and the densely distributed area of ​​the calibration points, so that the center of the gaze point is consistent with the center of the calibration point.

[0029] Furthermore, the gap is the Euclidean distance between the gaze point and the calibration point.

[0030] Compared with the prior art, the present invention has the following advantages:

[0031] The present invention measures the visual field of a person with visual field loss and performs initial calibration on a normal area through a perimeter and a grid, thereby ensuring the accuracy of subsequent detailed classification of the visual field area; the present invention can quickly and accurately obtain the visual field area of ​​a person with visual field loss through a custom grid without spending a lot of time; because the points of the present invention are evenly arranged, the visual field area of ​​the person with visual field loss can be quickly divided into a normal area, a distorted area and an invisible area by using a real gaze point distribution map, with high accuracy; the present invention performs gaze point calibration on different visual fields by region, performs normal mapping calibration on the normal visual field, performs triangulation and merging on the distorted visual field and then calibrates it, and at the same time uses a grid to check the calibration correctness of the distorted area.

[0032] The present invention can be used for gaze point calibration of people with visual field loss, can complete the task quickly and accurately, and has the characteristics of strong scalability and high robustness. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0034] Figure 1 It is a schematic diagram of the overall process of the present invention.

[0035] Figure 2 This is a comparison diagram of the visual field of people with normal visual field and people with visual field defect in the present invention. Among them, (a) is the visual field of normal people; (b) is the visual field of people with visual field defect.

[0036] Figure 3 It is a schematic diagram of the visual field partitioning of the present invention.

[0037] Figure 4 It is a schematic diagram of the field of view area calibration of the present invention. DETAILED DESCRIPTION

[0038] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0039] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0040] like Figure 1-Figure 4 As shown, the present invention provides a method for calibrating the sight line of human eyes in different regions for people with visual field impairment, comprising the following steps:

[0041] Step 1: Initial calibration; The visual field of the visual field loss group is measured by the perimeter and the grid to obtain the visible area I and the invisible area I, and the normal visual field in the visible area I is mapped and calibrated; the visible area I includes the visual field distortion area and the normal area; the visual field loss group is required to sit in front of the perimeter before use, with the lower jaw fixed on the jaw frame, and the eyes are fixed on the white fixed point at zero degrees on the arc of the perimeter. The sight mark (a small ball with a diameter of 3mm and a color of white) moves slowly from the periphery of the arc to the center. When the subject finds the sight mark, press the record button, and record the position of the sight mark on the arc at this time on the peripheral visual field table. Rotate the arc 30 degrees or 45 degrees and repeat the above steps until the arc rotates a circle. Finally, connect the points to get the visual field range of the eye. After obtaining the visual field area of ​​the visual field loss group through the perimeter, the distorted area is simply judged by the grid. The grid is a 10cm×10cm square with a 20-row×20-column table inside. The table needs to be placed 30cm from the horizon of the visual field loss group. The participant covers his left or right eye with his hand, and fixes his other eye on the black dot in the center of the table, and reports to the examiner whether there is any distorted part in the visual field and the specific location. The visual field is roughly divided into visible area and invisible area by the results of the perimeter and grid, and the visible area may include distorted visual field area and normal area.

[0042] In order to ensure the accuracy of subsequent field of view division and line of sight calibration, it is necessary to perform an initial calibration on the normal area based on the current field of view area division. All calibration processes mentioned in the present invention require measuring the characteristics of the user's eyes and using them together with a 3D eye model to calculate line of sight data. The model includes information such as the shape of different positions of the eyes, light refraction characteristics, and reflection characteristics. During the calibration process, first select the largest normal field of view area in the field of view result, and set calibration points at the four vertices and center point of the selected area. People with visual field defects are required to keep their eyes open and stare at the calibration points when the calibration points appear. The camera will collect eye information, integrate the eye feature data into the 3D eye model to calculate the gaze data before calibration, and compare multiple sets of coordinate data between the gaze point and the calibration point to fine-tune the parameters of the line of sight tracking method, thereby completing this calibration.

[0043] Step 2: Divide the visual field area; obtain the gaze point distribution map according to the predetermined acquisition method (there are densely arranged grids on the display device, and the visual field impaired people are required to look at the evenly arranged grid gaze points generated by the program in turn and report the grid where the gaze points are located.) (The visual field impaired people need to report the position of the current gaze point according to the grid information after being informed of the presence of the gaze point. If the gaze point is not observed, the record is empty; if the gaze point is observed, the position of the gaze point is recorded. If the observed position is the same as the actual one, it is a normal visual field; if the observed position is different from the actual one, it is a distorted visual field; if the position cannot be observed, it is an invisible visual field.) According to the gaze point distribution (the gaze point distribution map is an evenly distributed grid, and each grid records the actual gaze point distribution observed by the visual field impaired people. There may be normal visual fields, distorted visual fields, and invisible visual fields. Gaussian blurring the edges of the three areas can be converted into gaze point distribution), scan and merge the visual field area into normal visual fields, distorted visual fields, and invisible visual fields;

[0044] The real gaze point distribution map of the visual field impaired population is obtained through the set acquisition method, and the visual field of the visual field impaired population is divided in detail through the gaze point distribution map. After completing the simple calibration of the visual field, considering that the visual field of the visual field impaired population is relatively complex, for example, the normal area may contain a distorted area, and the boundary between the normal area and the distorted area is relatively blurred, it is not rigorous to use only the visual field meter to determine the visual field distribution. Therefore, this step considers obtaining the real gaze point distribution map of the visual field impaired population through the set acquisition method. The specific acquisition method is as follows: the visual field impaired population is required to sequentially gaze at the evenly arranged gaze points generated by the program. The program sets a total of 81 gaze points, including 9 rows of gaze points horizontally and 9 columns of gaze points vertically. There is a densely arranged grid on the display device. The visual field impaired population needs to report the position of the current gaze point according to the grid information after being informed of the appearance of the gaze point. If no gaze point is observed, the record is empty. After observing all the fixation points, we can get the fixation point information of people with visual field loss. We can divide them into three categories according to the arrangement rules of fixation points, namely normal visual field, distorted visual field and invisible visual field. The normal visual field appears as an orderly and uniform arrangement of fixation points in the fixation point distribution map; the distorted visual field appears as a more chaotic arrangement of fixation points in the fixation point distribution map; the invisible visual field appears as a missing fixation point in the fixation point distribution map.

[0045] According to whether the gaze points in the real world can be correctly observed, the visual field of people with visual field loss can be divided into three types: normal visual field, distorted visual field, and invisible visual field. The normal visual field is shown as an orderly and uniform arrangement of gaze points in the gaze point distribution map; the distorted visual field is shown as a relatively chaotic arrangement of gaze points in the gaze point distribution map; and the invisible visual field is shown as a missing gaze point in the gaze point distribution map. The distorted visual field and the invisible visual field can be marked by checking the distribution and arrangement of points. The unmarked part is the normal visual field. The specific technical solution is as follows:

[0046] Since the fixation points are 81 points with the same spacing, the real fixation point distribution map is divided into a grid, and the square grid can be divided into 81 square sub-areas. For normal people, there should be a fixation point in each square sub-area. For people with visual field defects, the abnormalities in the sub-areas are distorted visual fields and invisible visual fields.

[0047] After the gaze point distribution map is divided into grids, it is necessary to traverse each small square area to check whether there is a gaze point in the center of the square. If not, mark the area as invisible vision; if there is a gaze point but not in the center of the square, mark the area as distorted vision; if there is a gaze point in the center, mark the area as normal vision.

[0048] After completing the marking of each small area, the whole needs to be processed twice, and the visual field areas with the same mark that are connected together are corrected and merged, so as to completely divide the entire real gaze point distribution map into several rectangular blocks and complete the division of the overall visual field area. First, start searching from the first square of each row, add the first square to the current sequence, update the detection mark to the mark of the first square, and set the marking point to the first square of the row. If the mark of the square on the right is the same as the detection mark at this time, the square on the right is added to the sequence; if the mark of the square on the right is different from the detection mark at this time, merge the mark point to the current left rectangular block, update the detection mark to the mark of the right square, and update the marking point to the right rectangle; repeat the above process until the end of the row. After the single row processing is completed, the visual field area is merged. Starting from the combination block in the first row, get its length and mark it as used. If there is the same combination of the same length in the next row, mark it as used and merge it. If there is no combination of the same length, replace it with the next unused combination. Since the gaze point usually cannot cover the entire distortion area, the visual field distortion area will be expanded by one tenth on the original basis. Therefore, the outer layer of the selected visual field distortion area is part of the normal visual field area, and the inner layer is the visual field distortion area. Repeat the above steps to complete the division of the visual field area.

[0049] Step 3: Calibrate by region; define the invisible field of view as the field of view boundary and do not process it; calibrate the normal field of view and the distorted field of view separately; the normal field of view is calibrated by plane mapping; the distorted field of view is triangulated, and the affine transformation parameters of the local triangulation area are solved, and the triangular areas with roughly the same transformation matrix are merged, and the affine transformation parameters of the merged area are solved.

[0050] In the process of converting eye information to gaze point, the key is to define a coordinate system mapping function from the human eye plane coordinate system to the real plane coordinate system. This mapping function can convert the information of the pupil center position and the corneal reflection point into the coordinate information of the gaze position. Since the display calibration point and the plane where the gaze point is obtained are the same plane, it can be assumed that the point of sight is P and the pupil center is (e x ,e y ), the intersection of the sight lines in different directions is O, then the coordinates of point P can be expressed as P = O + λ · OP m , where P m =f(e x ,e y ), λ is the proportionality coefficient. If the coordinates of the line of sight intersection point O and the representation of the function f can be obtained, the corresponding landing point coordinates can be solved according to the human eye image. Due to the fixed plane, this method will not have parallel errors when calculating the landing point coordinates.

[0051] The present invention uses a polynomial with the highest order term being quadratic to implement the mapping rule. m =[x,y,z] T ,Right now:

[0052]

[0053] in β=[β x β y β z ] represents the mapping parameters. After obtaining multiple sets of data pairs, the generalized matrix is ​​used to solve the regression parameters. Although the present invention performs calibration on a fixed plane of the head-mounted device, considering that the user cannot completely keep the head still, the depth z is also calibrated to ensure the accuracy of the calibration.

[0054] In the calibration process of normal field of view, the coordinate transformation is linear. Therefore, you only need to follow the above steps to generate calibration points at the four vertices of the normal field of view. People with visual field defects are required to look at the calibration points normally. After the camera captures the eye image and obtains the offset data, you can adjust the program to compensate for the offset, thereby completing the line of sight calibration of normal field of view.

[0055] In the calibration process of the distorted visual field, although the visual field distortion does not have a relatively consistent regularity, the range of visual field defects is usually small. Therefore, the distorted visual field fixation area can be triangulated, and each triangular area can be identified as an affine transformation.

[0056] Affine transformation refers to a linear transformation of a vector space followed by a translation in geometry, thereby transforming it into another vector space. Assume that in the original vector space, k = (x, y), j = (x ′ ,y ′ ), then the conversion between the two vector spaces can be realized through the parameter matrix M, the formula is as follows:

[0057] [x ′ ,y ′ ]=M[x,y]

[0058] If the parameter matrices obtained for adjacent triangular regions are roughly the same, it means that the transformation rules of the two triangular regions are consistent, so the two triangular regions can be merged, and the same applies to other regions.

[0059] When realizing the sight calibration of the visual distortion area, it is first necessary to perform Delaunay triangulation on the visual distortion area and the corresponding real area, where the visual distortion area corresponds to part of the area on the normal screen. The construction algorithm of Delaunay triangulation is implemented by point-by-point insertion. Whenever a new vertex is inserted, connecting it with the nearby old vertices will add some new edges. We only need to use the flip edge algorithm to correct the "bad edges" in it, and then we can "repair" it to Delaunay triangulation. Repeat the above steps to complete the triangulation of the area. Secondly, it is necessary to solve the affine transformation of each triangulation, and complete the calculation of the parameter matrix through the relationship between the calibration point and the fixation point in the triangulation. Then, it is determined whether the triangular area around each triangulation has the same parameter matrix as itself. If so, different triangulations can be merged, and the merged visual distortion area has roughly the same mapping rules; if the triangulation cannot be merged with other areas, the triangular area has a separate mapping rule.

[0060] When performing sight calibration, it is necessary to try to ensure that the calibration points appear evenly in all triangulated and merged areas. If it is a merged area, the gaze point appears at the vertex and center point; if it is an unmerged triangulation, the gaze point appears at the center of the triangle. Similar to the calibration of normal areas, people with visual field defects need to keep their eyes open and look at the calibration points when the calibration points appear. During this process, the affine matrix of each area is added to the calibration process in reverse. Until the distorted area is calibrated. After completing this calibration, in order to ensure the accuracy of the calibration, it is necessary to compare the gaze point and the calibration point of the distorted visual field area again. If the difference (the difference is the Euclidean distance between the gaze point and the calibration point) is within an acceptable range, the area is considered to be calibrated, otherwise it is recalibrated according to the above process.

[0061] Preferably, when implementing the method of the present invention, the visual field tracking of people with visual field loss can be achieved by remote visual field tracking or head-mounted visual field tracking. Specifically:

[0062] In telemetric gaze tracking, the interactive device for people with visual field impairment is the screen in front of them. People with visual field impairment need to look at the points on the screen for visual field calibration, and the calculated gaze points also appear on the screen. When looking at the screen, the camera above the screen will collect eye data for calculation. In order to ensure the quality of eye data collection, additional light sources are often installed around the camera. The coordinate transformation of telemetric gaze tracking involves the camera coordinate system, the human eye coordinate system, and the screen coordinate system. First, the camera collects human eye data, which needs to be converted from the camera coordinate system to the human eye coordinate system, and then mapped from the human eye coordinate system to the screen coordinate system, so that the gaze point can be displayed on the screen.

[0063] In head-mounted gaze tracking, the interactive device for people with impaired visual field is a gaze tracking instrument worn on the head. At the same time, an external screen is needed to monitor the internal situation of the head-mounted device. People with impaired visual field need to look at the gaze point that appears in the head-mounted device for visual field calibration and gaze tracking. The tester needs to look at the external screen to observe the visual field of the person with impaired visual field. Since head-mounted devices are often highly integrated, head-mounted gaze calibration does not require additional light sources and is simpler to arrange. Since the position of the camera and the screen is almost the same, the coordinate transformation of head-mounted gaze tracking involves the screen coordinate system and the human eye coordinate system. The camera collects human eye data and maps the coordinates to the screen to complete the display of the gaze tracking gaze point.

[0064] Compared with head-mounted gaze tracking, head-mounted gaze tracking is easier to implement, but it is more burdensome for participants and difficult to carry out for a long time; remote gaze tracking needs to consider the placement of cameras and light sources, and the conversion of coordinates is more difficult, but participants only need to look at the screen to realize all functions, and no additional equipment is required, which is suitable for long-term gaze tracking. Both have their own advantages, and there is no strict distinction between good and bad. You only need to choose according to the actual situation during use.

[0065] Step 4: After calibration is completed, the gaze point and calibration point are compared again in the distorted area. If the difference is within the threshold range, the calibration is considered complete; if the difference is greater than the threshold range, triangulation is performed again, and the centers of the three newly generated triangles are used as new calibration points until the difference is within the threshold range. The threshold range is 1% of the size of the long side of the screen. The difference is the Euclidean distance between the gaze point and the calibration point.

[0066] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0067] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0068] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0069] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0070] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0071] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0072] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for calibrating the sight line of the human eye in different regions for people with visual field impairment, characterized in that: The following steps are involved: Step 1: Initial calibration; The visual field of the visual field-impaired population is measured by using a perimeter and a grid to obtain a visible area I and an invisible area I, and a normal visual field in the visible area I is mapped and calibrated; the visible area I includes a visual field distortion area and a normal area; Step 2: Divide the visual field area; obtain the gaze point distribution map according to the predetermined acquisition method, and scan and merge the visual field area into normal visual field, distorted visual field and invisible visual field according to the distribution of gaze points; Step 3: Calibrate by region; Define the invisible field of view as the field of view boundary and do not process it; The normal field of view and the distorted field of view are calibrated respectively; the normal field of view is calibrated by plane mapping; the distorted field of view is triangulated, and the affine transformation parameters of the local triangulation area are solved, and the triangular areas with roughly the same transformation matrix are merged to solve the merged affine transformation parameters; Step 4: After calibration is completed, the gaze point and calibration point of the distorted area are compared again. If the difference is within the threshold range, the calibration is considered complete. If the difference is greater than the threshold range, triangulation is performed again, and the centers of the three newly generated triangles are used as new calibration points until the difference is within the threshold range.

2. The method for calibrating the sight line of the human eye in different regions for people with visual field impairment according to claim 1, characterized in that: During the initial calibration process, firstly, the largest normal visual field area is selected in the visual field result, and calibration points are set at the four vertices and the center point of the selected area respectively; The user is required to keep their eyes open and stare at the calibration points when they appear. The camera is used to collect eye information, and the eye information data is integrated into the 3D eye model to calculate the gaze data before calibration. The parameters of the gaze tracking method are fine-tuned by comparing multiple sets of coordinate data of the gaze points with those of the calibration points.

3. A method for regional human eye sight calibration for people with visual field impairment according to claim 1 or 2, characterized in that: During the calibration process of the initial calibration, it is necessary to measure the features of the user's eyes and calculate the sight line data together with the features of the user's eyes and the 3D eye model.

4. The method for calibrating the sight line of the human eye in different regions for people with visual field impairment according to claim 1, characterized in that: The field of view area division process comprises the following steps: Step 21: The user sequentially gazes at evenly arranged gaze points generated by the program; Step 22: The user reports the location of the current fixation point according to the grid information. If no fixation point is observed, the record is empty. The fixation point information of the field-deficient population is obtained by observing all fixation points. Step 23: Divide the gaze point information of the field-deficient population into normal visual field, distorted visual field and invisible visual field according to the arrangement rules of the gaze points.

5. The method for calibrating the sight line of the human eye in different regions for people with visual field impairment according to claim 4, characterized in that: The normal visual field is shown as an orderly and uniform arrangement of gaze points in the gaze point distribution map; the distorted visual field is shown as a chaotic arrangement of gaze points in the gaze point distribution map; and the invisible visual field is shown as a missing gaze point in the gaze point distribution map.

6. The method for calibrating the sight line of the human eye in different regions for people with visual field impairment according to claim 1, characterized in that: In step 3, for the normal field of view correction process, the normal field of view mapping is achieved through a polynomial; let P m =[x,y,z] T ,Right now: in, β=[β x β y β z ] represents mapping parameters; P m Indicates the coordinate position; x, y, z represent the coordinates of the x, y, z axes respectively; generate calibration points at the four vertices of the normal field of view, and require the user to look at the calibration points normally. After the camera captures the eye image and obtains the offset data, the program can be adjusted to compensate for the offset, thereby completing the line of sight calibration of the normal field of view.

7. The method for calibrating the sight line of the human eye in different regions for people with visual field impairment according to claim 1, characterized in that: In the step 3, for the calibration process of the distorted field of view, a vector space is linearly transformed in geometry by an affine transformation and then a translation is performed, thereby transforming into another vector space; Assuming that k = (x, y), j = (x', y') in the original vector space, the conversion between the two vector spaces is realized through the parameter matrix M, and the formula is: [x ′ ,y ′ ]=M[x,y]; If the parameter matrices obtained for adjacent triangular regions are roughly the same, it means that the transformation rules of the two triangular regions are consistent, and the two triangular regions are merged.

8. The method for calibrating the sight line of the human eye in different regions for people with visual field impairment according to claim 1, characterized in that: The threshold range is one percent of the long side of the screen.

9. The method for calibrating the sight line of the human eye in different regions for people with visual field impairment according to claim 2, characterized in that: The fine-tuning refers to automatically adjusting the coordinate parameters generated by the model when calculating the overall center coordinates of the densely distributed area of ​​the gaze points and the densely distributed area of ​​the calibration points, so that the center of the gaze point is consistent with the center of the calibration point.

10. The method for calibrating the sight line of the human eye in different regions for people with visual field impairment according to claim 1, characterized in that: The gap is the Euclidean distance between the fixation point and the calibration point.

Images