CN112926536A - Deformed pupil positioning method, device and equipment - Google Patents
Deformed pupil positioning method, device and equipment Download PDFInfo
- Publication number
- CN112926536A CN112926536A CN202110369394.8A CN202110369394A CN112926536A CN 112926536 A CN112926536 A CN 112926536A CN 202110369394 A CN202110369394 A CN 202110369394A CN 112926536 A CN112926536 A CN 112926536A
- Authority
- CN
- China
- Prior art keywords
- pupil
- parameter
- positioning
- value
- deformed
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 64
- 208000022749 pupil disease Diseases 0.000 title claims abstract description 50
- 210000001747 pupil Anatomy 0.000 claims abstract description 154
- 230000008569 process Effects 0.000 claims description 28
- 230000015654 memory Effects 0.000 claims description 25
- 238000004590 computer program Methods 0.000 claims description 19
- 230000003247 decreasing effect Effects 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000013500 data storage Methods 0.000 claims description 4
- 230000004424 eye movement Effects 0.000 claims description 4
- 210000000720 eyelash Anatomy 0.000 abstract description 5
- 210000000744 eyelid Anatomy 0.000 abstract description 4
- 210000004209 hair Anatomy 0.000 abstract description 3
- 238000004422 calculation algorithm Methods 0.000 description 26
- 230000006870 function Effects 0.000 description 22
- 238000012545 processing Methods 0.000 description 16
- 238000005457 optimization Methods 0.000 description 10
- 239000013598 vector Substances 0.000 description 9
- 238000004458 analytical method Methods 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 238000012360 testing method Methods 0.000 description 4
- 238000013461 design Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 230000000670 limiting effect Effects 0.000 description 3
- 230000004807 localization Effects 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 238000009795 derivation Methods 0.000 description 2
- 235000019800 disodium phosphate Nutrition 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 230000036961 partial effect Effects 0.000 description 2
- 230000002829 reductive effect Effects 0.000 description 2
- 238000003491 array Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000010845 search algorithm Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/18—Eye characteristics, e.g. of the iris
- G06V40/193—Preprocessing; Feature extraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- Ophthalmology & Optometry (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Human Computer Interaction (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a deformed pupil positioning method, a device and equipment, and the idea of the invention is to lock the approximate range of pupils in an eye image through preliminary positioning, and on the basis of the range, aiming at the characteristic that the deformed pupil positioning has higher space complexity and time complexity, and combining the pixel gray scale of the eye image and a preset differential evolution strategy, global evolutionary parameter search is carried out in the preliminary positioning result to accelerate the obtaining of the optimal solution of pupil shape parameters. The method can efficiently search the position of the deformed pupil in the input eye moving image, and particularly can obviously improve the speed and the precision of pupil positioning aiming at the condition of the existence of uneven light spots, eyelashes, hair, eyelids, spectacle frames and other interference factors.
Description
Technical Field
The invention relates to the technical field of sight line tracking, in particular to a deformed pupil positioning method, a deformed pupil positioning device and equipment.
Background
In different gaze tracking application scenarios, a target for gaze tracking, that is, a pupil, may be deformed to some extent during imaging due to configuration differences of a tracking device or changes of an object to be detected itself, and in most cases, a circular pupil may be changed into an elliptical pupil.
The current positioning technology for the deformed pupil mainly depends on the premise of accurately extracting the edge points of the deformed pupil in the image, that is, the accuracy of pupil positioning is directly influenced by the good and bad extracted object edge points. Moreover, in the process of extracting the edge points, a large amount of search is performed in a multidimensional parameter space, so that the realization efficiency is unsatisfactory, and especially when an eye moving image with interference factors is confronted, for example, the eyes are shielded by light spots, eyelashes, eyelids and the like, a large amount of interference noise is introduced, so that the complexity of realizing the deformed pupil positioning is increased, and the problems that the time consumption for solving the search process is obvious, the positioning fails and the like can also occur are caused.
Disclosure of Invention
In view of the above, the present invention aims to provide a deformed pupil positioning method, apparatus and device, and accordingly provides a computer data storage medium and a computer program product, which mainly solve the problems that the existing deformed pupil positioning scheme depends on the extraction accuracy of pupil edge pixel points, so that the search efficiency is low and the positioning is easy to be interfered, resulting in positioning failure.
The technical scheme adopted by the invention is as follows:
in a first aspect, the present invention provides a deformed pupil positioning method, which includes:
performing pupil preliminary positioning on the obtained to-be-detected eye moving image to obtain the range of pupil parameters; wherein the pupil parameters include of the pupil: the central abscissa, the central ordinate, the major axis, the major and minor axis proportions and the deflection angle;
acquiring the gray value of the pixel point meeting the pupil parameter in the eye moving image to be detected;
and searching all pixel points which accord with the range according to the gray value and a preset differential evolution strategy and determining the target position of the pupil in the eye image to be detected.
In at least one possible implementation manner, the searching, according to the gray value and a preset differential evolution policy, all the pixel points that meet the range includes:
based on a preset ellipse difference operator, solving gray level difference values of all pixel points in the range;
searching an extreme value of the gray level difference value, and determining a target pupil parameter according to the difference extreme value.
In at least one possible implementation manner, the searching, according to the gray value and a preset differential evolution policy, all the pixel points that meet the range specifically includes:
constructing a parameter population based on the range of the pupil parameter, wherein the parameter population comprises a plurality of parameter individuals;
expanding the parameter individuals;
calculating gray level difference values of all pixel points in the range by combining the parameter individuals before and after expansion and the ellipse difference operator;
comparing the calculation results of the parameter individuals before and after expansion, updating the parameter population by using the parameter individuals corresponding to the extreme value of the gray level difference value, and evolving to obtain a next generation parameter population;
and searching the pupil parameters according to the process until the preset evolution condition is met, and outputting the target pupil parameters.
In at least one possible implementation manner, expanding the parameter individuals according to the description includes:
carrying out variation on the parameter individuals according to a preset scaling factor to obtain variation individuals corresponding to each parameter individual;
and generating a plurality of intermediate individuals according to the parameter individuals, the variant individuals and a preset cross probability factor.
In at least one possible implementation manner, the values of the scaling factors are in a generation-by-generation decreasing trend along with the increase of the population evolution generation, and/or the values of the cross probability factors are in a generation-by-generation increasing trend along with the increase of the population evolution generation.
In a second aspect, the present invention provides a deformed pupil positioning device, comprising:
the preliminary positioning module is used for carrying out preliminary pupil positioning on the acquired to-be-detected eye movement image to obtain the range of pupil parameters; wherein the pupil parameters include of the pupil: the central abscissa, the central ordinate, the major axis, the major and minor axis proportions and the deflection angle;
the pixel gray level acquisition module is used for acquiring the gray level value of the pixel point meeting the pupil parameter in the eye moving image to be detected;
and the pupil parameter searching module is used for searching all pixel points in the range according to the gray value and a preset differential evolution strategy and determining the target position of the pupil in the eye image to be detected.
In at least one possible implementation manner, the pupil parameter search module includes:
the pupil parameter solving submodule is used for solving a gray level difference value of all pixel points in the range based on a preset ellipse difference operator;
and the target parameter determining submodule is used for searching an extreme value of the gray level difference value and determining a target pupil parameter according to the difference extreme value.
In at least one possible implementation manner, the pupil parameter search module specifically includes:
the population initialization unit is used for constructing a parameter population based on the range of the pupil parameter, and the parameter population comprises a plurality of parameter individuals;
the parameter individual expanding unit is used for expanding the parameter individuals;
the parameter solving individual unit is used for calculating gray level difference values of all pixel points in the range by combining the parameter individuals before and after expansion and the ellipse difference operator;
the parameter individual selection unit has calculation results of parameter individuals before and after full-bodied comparison expansion, updates the parameter population by using the parameter individual corresponding to the extreme value of the gray level difference value, and evolves to obtain a next generation parameter population;
and the evolution termination unit is used for searching the pupil parameters according to the process until the preset evolution conditions are met and outputting the target pupil parameters.
In at least one possible implementation manner, the parameter individual expansion unit includes:
the individual variation component is used for performing variation on the parameter individuals according to a preset scaling factor to obtain variation individuals corresponding to each parameter individual;
and the individual crossing component is used for generating a plurality of intermediate individuals according to the parameter individuals, the variant individuals and a preset crossing probability factor.
In at least one possible implementation manner, values of scaling factors in the differential evolution strategy are preset to be in a generation-by-generation decreasing trend along with the increase of population evolution generations; and/or the value of the cross probability factor in the differential evolution strategy is preset to be in a generation-by-generation increasing trend along with the increase of population evolution generations.
In a third aspect, the present invention provides an electronic device, comprising:
one or more processors, memory which may employ a non-volatile storage medium, and one or more computer programs stored in the memory, the one or more computer programs comprising instructions which, when executed by the apparatus, cause the apparatus to perform the method as in the first aspect or any possible implementation of the first aspect.
In a fourth aspect, the present invention provides a computer data storage medium having a computer program stored thereon, which, when run on a computer, causes the computer to perform at least the method as described in the first aspect or any of its possible implementations.
In a fifth aspect, the present invention also provides a computer program product for performing at least the method of the first aspect or any of its possible implementations, when the computer program product is executed by a computer.
In at least one possible implementation manner of the fifth aspect, the relevant program related to the product may be stored in whole or in part on a memory packaged with the processor, or may be stored in part or in whole on a storage medium not packaged with the processor.
The method is characterized in that the approximate range of the pupil in the eye image is locked through primary positioning, and on the basis of the range, aiming at the characteristic that deformed pupil positioning has higher space complexity and time complexity, the global evolutionary parameter search is carried out in a primary positioning result by combining the pixel gray scale of the eye image and a preset differential evolution strategy so as to accelerate the obtaining of the optimal solution of the pupil shape parameter. The method can efficiently search the position of the deformed pupil in the input eye moving image, and particularly can obviously improve the speed and the precision of pupil positioning aiming at the condition of the existence of uneven light spots, eyelashes, hair, eyelids, spectacle frames and other interference factors.
Furthermore, an ellipse solving operator is preset for the differential evolution strategy, namely, an objective function for solving the pupil parameter is set for a parameter selection link in the differential evolution step, so that the robustness and the accuracy of the positioning solving process are improved in an auxiliary manner.
Further, in some preferred embodiments, the present invention optimizes some control factors of the differential evolution strategy, so that global optimality, convergence accuracy, convergence speed, etc. of the deformed pupil positioning are all significantly improved.
Drawings
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be further described with reference to the accompanying drawings, in which:
fig. 1 is a flowchart of an embodiment of a deformed pupil positioning method according to the present invention;
FIG. 2 is a flowchart of an embodiment of searching pupil parameters according to a differential evolution strategy provided in the present invention;
fig. 3 is a schematic diagram of an embodiment of a deformed pupil positioning device provided in the present invention;
fig. 4 is a schematic diagram of an embodiment of an electronic device provided in the present invention.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are illustrative only and should not be construed as limiting the invention.
Before the specific scheme of the present invention is developed, aiming at the aforementioned requirements for considering both the parameter search efficiency and avoiding the interference influence, the present invention combines the conventional ellipse fitting mode with the deformed pupil scheme for analysis and study, and of course, this is also the derivation process forming the inventive concept of the present invention. By the analysis of the inventor, the current three main strategies for deformed pupil positioning are as follows: and the algorithm of fitting the ellipse, the algorithm of least square fitting and the algorithm of random sampling consistency based on Hough transformation have advantages and disadvantages respectively. Specifically, the method comprises the following steps:
(1) hough transform fitting ellipse algorithm
The algorithm has the main advantages that the algorithm is relatively insensitive to partial defects of ellipses, noise and the like, and single requirement of detection precision can be guaranteed, but as the deformed pupils have 5 undetermined parameters, namely, accumulated voting needs to be carried out in a 5-dimensional parameter space when the elliptical pupils are detected, the time and space complexity of the algorithm is very high, and the fitting efficiency can not completely meet the requirement of the current real-time online application scene.
(2) Least squares fitting algorithm
The algorithm is mainly based on the principle of minimizing the mean square error, and can achieve higher fitting precision when the input condition that the pupil edge point is easy to detect is faced. However, in practical applications, the input eye image is often affected by the interference points in the process of detecting the pupil edge points, and especially under the condition that the interference on eyelashes, light spots and the like is serious, the elliptical pupil fitted in this way has a larger deviation than the real result.
(3) Algorithm for random sampling consistency
The probability algorithm may be superior to the former in fitting speed, but it is easy to introduce invalid sampling at a still high probability in the input image with large disturbance, so that both the amount of computation and the memory requirement increase, and may eventually cause the algorithm to have difficulty locating the target pupil with acceptable time or memory space occupation.
Through analysis, the algorithm cannot take efficiency and precision into consideration, so that the invention provides a deformed pupil positioning scheme which does not rely on pupil edge point extraction any more. Specifically, the present invention provides an embodiment of at least one deformed pupil positioning method, as shown in fig. 1, which specifically includes:
and step S1, performing pupil preliminary positioning on the obtained to-be-detected eye movement image to obtain the range of pupil parameters.
The pupil is circular, and most of the deformed structures presented in the eye images are ellipses, so the pupil parameters considered by the invention are considered according to the ellipse parameters, and specifically, the pupil parameters mainly comprise the central abscissa of the pupil, the central ordinate of the pupil, the long axis of the pupil, the proportion of the long axis and the short axis of the pupil, and the deflection angle of the pupil.
In other words, the need for deformed pupil localization translates into the task of finding a globally optimal solution in a 5-dimensional parameter space, which is an extremely large and complex search task, as will be appreciated. Further, according to analysis, since the pupil in the eye image has a more obvious image characteristic, in the embodiment, it is proposed that the approximate range of the pupil in the eye image can be preliminarily locked by a preliminary positioning manner, so as to accelerate the processing speed of the subsequent accurate positioning. The preliminary positioning mentioned here can be understood as coarse positioning, and in actual operation, it can be implemented by, but not limited to, using a priori knowledge, conventional binarization, radial symmetry, and other mature strategies.
For example, in some implementation references, the rough positioning strategy may be used to obtain the rough position of the pupil in the eye image, for example, for the eye image to be measured (or the input eye image) with the resolution of 768 × 576, the rough positioning process may be used to obtain the rough search interval of the pupil parameter: the variation range of the transverse and longitudinal coordinates of the pupil center is 100 pixels by 100 pixels; the long axis size is between 20-120 pixels; the scale factor of the short axis and the long axis is 0.2-1.0; the deflection angle is between 0 and 180 degrees. In addition, assuming that in this example, the search step size of the center position and the long axis is 1 pixel, the scale factor step size is 0.01, and the deflection angle step size is 1 degree, then after the preliminary positioning, the size of the obtained traversal space is:
Total＝Δx*Δy*Δa*Δε*Δθ
＝100*100*100*80*180＝1.44*1010
wherein Δ x, Δ y, Δ a, Δ ∈, Δ θ are the horizontal and vertical coordinates of the center of the deformed pupil, the long axis, the scale factor, and the variation of the deflection angle, respectively.
And step S2, acquiring the gray value of the pixel point meeting the pupil parameter in the eye moving image to be detected.
The invention has the advantages that the target search is carried out on the image, so the attribute characteristics of the image are required to be used, the gray value of the pixel point of the eye image is used as the investigation basis, the technology for acquiring the gray value of the pixel belongs to the mature technology in the field, and the invention can point out that the camera equipment adopted in the relevant application scene can conveniently acquire the gray attribute of the pixel, for example, an infrared camera used by a head-mounted sight tracking system can directly capture the eye gray image, so the invention can not limit the implementation mode of the step.
And step S3, searching all pixel points in the range according to the gray value and a preset differential evolution strategy, and determining the target position of the pupil in the eye image to be detected.
As mentioned above, even though the coarse positioning process is performed to narrow the global search range of the input image to a smaller range, the search range is relatively limited to 1.44 × 10 in the case of the above example10If the conventional point-by-point traversal search algorithm is adopted, the processing efficiency still cannot be guaranteed. Based on this, the embodiment introduces a differential evolution strategy to perform efficient iterative search on the result after the coarse positioning, so as to improve the efficiency of pupil positioning.
In order to ensure the precision, the method is a preset ellipse difference operator in actual operation, the gray level difference value of the ellipse circumference is obtained for all the pixel points in the range, and therefore after the extreme value of the gray level difference value is searched, the target pupil parameter can be determined according to the difference extreme value.
Based on this, in at least one preferred embodiment of the present invention, a better solution objective function is proposed, namely, the pupil parameters are obtained by using the following calculus operator for solving the ellipse:
wherein
I (x, y) is the gray value of the eye image to be measured at the coordinate (x, y), a is the major axis corresponding to the ellipse, epsilon is the scale factor of the minor axis and the major axis of the ellipse, theta is the deflection angle of the ellipse, (x)c,yc) Is the central coordinate of an ellipse, Gσ(a) The standard deviation is a gaussian function with σ, and specifically can play a role of smoothing filtering, and 2 π a ε +4a (1- ε) is the perimeter of the ellipse, that is, the embodiment takes the above formula as the objective function to deform the ellipse by differential evolutionAnd the pupil carries out parameter solution. Specifically, the ellipse solver can be used to assume the center of the circle as (x)c,yc) And on an ellipse circle ds with the major axis a, the minor axis a epsilon and the deflection angle theta, integrating (and performing normalization processing) the gray values of all the pixel points in a small range after the coarse positioning, then obtaining a gray difference value, and taking parameters (x, y, a, epsilon and theta) corresponding to the difference extreme value as parameters for positioning the target pupil position such as the center of the pupil, the major axis and the like.
In other words, the effect of the above-mentioned overall ellipse difference operator can be understood as an ellipse edge detector based on scale σ blurring, which does not need to extract ellipse edge pixel points as in the conventional ellipse fitting method, but rather obtains a global parameter space I (x) after coarse positioningc,ycAnd a, epsilon and theta) and solving an optimal solution, and the method can achieve single-pixel precision in the process of fuzzily searching the pupil edge, thereby effectively avoiding interference factors in the eye moving image to be detected.
That is, with the complexity and diversity of the solution problem, it is difficult for the conventional deterministic optimization algorithm in the art to solve the optimization problem of searching the gray difference maximum on the ellipse circumference in the 5-dimensional space, so the invention combines the ellipse solution formula and the differential evolution strategy to realize the positioning of the elliptical pupil.
Through comparison and analysis in the design stage, the differential evolution strategy is more expected to be expressed on the aspect of solving the problem of how to consider the positioning efficiency and the positioning precision. Specifically, the differential evolution strategy is a greedy genetic algorithm with an optimization-preserving idea based on real number coding, has strong global search capability, and has the characteristics of high convergence speed, high efficiency, strong robustness and the like. In actual operation, a differential evolution algorithm (DE) can carry out evolution based on the difference of parameter populations, can memorize the optimal solution of individuals and share information in the population, and realizes the solution of the optimization problem through cooperation and competition among individuals in the population.
Referring to fig. 2, the present invention provides a parameter search scheme preferably combining the idea of differential evolution, which may include the following steps:
step S31, constructing a parameter population based on the range of the pupil parameter, wherein the parameter population comprises a plurality of parameter individuals;
step S32, expanding the parameter individuals;
step S33, calculating gray level difference values of all pixel points in the range by combining the parameter individuals before and after expansion and the ellipse difference operator;
step S34, comparing the calculation results of the parameter individuals before and after expansion, updating the parameter population by using the parameter individuals corresponding to the extreme value of the gray level difference value, and evolving to obtain a next generation parameter population;
and step S35, searching the pupil parameters according to the process until the preset evolution conditions are met, and outputting the target pupil parameters.
In order to fully utilize the cooperation and competition relationship among individuals in the parameter population, the operation of expanding the parameter individuals can be embodied as follows: on one hand, the parameter individuals are subjected to variation according to a preset scaling factor, and variation individuals corresponding to the parameter individuals are obtained; and on the other hand, generating a plurality of intermediate individuals according to the parameter individuals, the variant individuals and a preset cross probability factor. In other words, before differential evolution and preferential operation is performed on the ellipse circumference, the parameter individuals in the parameter population of the current generation are fully expanded, the specific expansion mode combining variation and intersection is a preferable illustration of the invention, and the expansion of the parameter individuals by adopting other concepts in other embodiments is not excluded.
The foregoing concept is further explained with reference to specific examples:
(1) and initializing the population parameters based on the elliptical pupil positioning scene.
Selecting individual parameter vector of the population as [ a, epsilon, theta, x [ ]c,yc]Wherein the physical meanings of the individual parameters are as follows: a represents the major axis, ε represents the minor and major axis scale factors, θ represents the deflection angle, xc,ycRepresents the ellipse circumference center coordinates; thereby, canEach generation of population is constructed using NP real-valued parameter vectors of dimension 5, expressed as:
each individual parameter in the population is:
1,2, …, NP, wherein: i is the serial number of the individual in the population; g is evolution algebra; NP is the population size; the lower corner mark 12345 is used to identify 5 parameters associated with a deformed pupil, the different parameter dimensions being identified hereafter by the symbol j. It should be noted that, in the initial stage, the maximum evolutionary algebra G can be presetmaxAnd scaling factor and cross probability factor values for individual expansion (e.g., F ∈ [0,2 ]]，CR∈[0,1]And one of the fixed values may be selected in practical applications).
Generally, it can be assumed that all randomly initialized populations obey uniform probability distribution, so for a deformed pupil positioning scene, population parameter initialization can also be randomly selected, and the specific operations are that parameter value intervals obtained from the coarse positioning result (long axis: 20 < a < 120, scale factors of short axis and long axis: 0.2 < epsilon < 1.0, deflection angle: 0 < theta < 180, central coordinate: 1 < x)c＜100，1＜yc< 100)), and NP real-number parameter vectors with the dimension of 5 are randomly selected from the parameter space with five dimensions as an initial first generation population.
(2) Individual variations for the elliptical pupil localization parameters.
As mentioned above in the preferred example, individual vectors can be applied to each elliptical pupil parameter in the current generation population
Performing variation operation to obtain elliptical pupil parameter variation individual vector corresponding to the variation operation
Specifically, it can be obtained by the following formula:
wherein the randomly selected sequence number r1,r2,r3E {1,2, …, NP } are different from each other and from i;
can be considered as a parent basis vector of the pupil parameters,
can be used as a pupil parameter parent differential vector; here, F is a scaling factor of the elliptical pupil parameter, and its function is to control the utilization ratio of the current difference.
(3) Individual crossings for elliptical pupil location parameters.
In order to increase the diversity of individual parameters, the embodiment introduces a crossover operation, specifically, the crossover operation is performed according to the component of the individual vector of each elliptical pupil parameter, and the specific implementation process can refer to the following steps, firstly generating a random integer randn (i), and then generating a variant individual of the pupil parameter according to the generated random integer randn (i)
And individual parameters in the parameter population
Intermediate individuals that generate elliptical pupil parameters:
more preferably, to ensure the progressiveness of the elliptical pupil parameter, the intermediate individuals in the elliptical pupil parameter can be made to be selected randomly
At least one individual capable of passing through the above mutation
To contribute and the other bits may be selected according to a cross probability factor, the optimization logic being as follows.
Wherein rand (j) represents a uniformly distributed random number between [0,1 ]; CR is a cross probability factor ranging between [0,1 ]; randn (i) is a random quantity between the elliptical pupil parameters {1,2,3,4,5 }.
(4) Elliptical pupil parameters are selected individually.
After the diversity of the population parameters is expanded through the foregoing process, the intermediate individuals of the pupil parameters can be generated by the above-mentioned elliptic differential operator
And individual parameters in the initial population of pupil parameters
The objective function of (1).
In practical operation, in order to facilitate the individual selection of the link, the extremum solving process of the partial differential equation of the elliptic objective function (elliptic difference operator) mentioned above is changed as follows, i.e. the solving problem of the extremum is transformed into the solving problem of the extremum in the five-dimensional parameter space I (x)c,ycA, ε, θ) solving the problem of the maximum:
wherein σ is the interval between two ellipses, so that two differential solution results can be obtained by using the parameter individuals before and after expansion (the parameter individuals and the intermediate individuals in this example), and further, the two differential solution results can be obtainedComparing the two populations to select the individual with better objective function value as the individual of new elliptical pupil parameter population
Namely:
wherein the content of the first and second substances,
i.e. the optimized objective function in this example.
(5) Termination test of the elliptical pupil parameters.
In the evolution process, if the G generation population X of the elliptic pupil parameterGIf the preset termination condition is met, outputting the target optimal solution, otherwise, repeating the operation in the above manner until the termination condition is reached, wherein the termination condition mentioned here can preset TmaxOr the maximum number of iterations G set in the initialization process mentioned above is reached as the threshold valuemax。
Based on the foregoing embodiments and their preferred solutions, the present invention further considers that, although the DE algorithm can meet the specific requirements of the present invention, the conventional DE algorithm itself has some disadvantages, such as:
(1) easily get into local optimum;
(2) the conventional DE sets the variation factor (scaling factor) F and the crossover probability factor CR therein mostly to fixed values.
Therefore, in some preferred embodiments, the invention performs optimization and improvement on the individual parameter expansion link in the expansion process, so as to achieve the purpose of further improving the search speed while ensuring the maximum value to be searched.
Specifically, the value of the scaling factor F may be gradually decreased with the increase of population evolution algebra, and/or the value of the cross probability factor CR may be gradually increased with the increase of population evolution algebra, that is, the individual expansion link in the differential evolution process is dynamically adjusted, instead of using the initially preset fixed standard.
The following description is given by way of a specific but non-limiting example of implementation. In the deformed pupil positioning process concerned by the invention, one of the expansion means related to the optimal differential evolution strategy is that the scaling factor F influences the convergence and the convergence speed of the search process.
In the individual variation process, when the F value is smaller, the population target individual aiming at the elliptic pupil parameter
The variation degree of the pupil parameter is smaller, so that the convergence speed of the pupil parameter is higher, but if F is too small, the pupil parameter is easy to fall into local optimum, namely, the final convergence is not an optimum solution; when the value of F is larger, the individual variation of the original parameters of the population for the elliptical pupil parameters is larger, and although the optimal solution is converged in the elliptical pupil parameter space, the convergence efficiency is lower.
Aiming at the contradiction points analyzed by the method, the method can consider that the diversity of the elliptical pupil parameter population is firstly kept in the initial stage of evolution search, then the convergence of the elliptical pupil parameters is gradually accelerated, and particularly, a dynamic scaling factor can be designed
Where G represents the current evolution algebra, GmaxRepresents a preset maximum number of optimization generations, [ F ]min,Fmax]The kernel function of the optimized scaling factor is the value range of the scaling factor determined based on the application scene
It can thus be understood that the optimized scaling factor is a decreasing function based on the population evolution algebra g of the elliptical pupil parameter. When g is 0, F is Fmax(ii) a When G ═ GmaxWhen F is equal to FminTherefore, in the early stage of differential evolution search, the scaling factor F is relatively large, the diversity of the elliptical parameter space population in the stage can be ensured, and the subsequent F is reduced generation by generation along with the increase of the evolution algebra, so that the convergence of the elliptical pupil parameter space can be accelerated in the later stage of evolution search.
The following description is given by way of a specific but non-limiting example of implementation. In the deformed pupil positioning process concerned by the invention, one of the expansion means related to the optimal differential evolution strategy, the cross probability factor CR also influences the convergence and the convergence speed of the search process.
In the process of cross expansion of individuals, if the CR value is larger, the variant individuals in the elliptic pupil parameter population
Intermediate individuals of population with elliptic pupil parameters
The larger the contribution of (c), and when CR is equal to 1, the two are equal, so that the local search and convergence speed of the elliptical pupil parameter can be increased. Otherwise, if the CR value is smaller, the target individual in the elliptic pupil parameter population is
For intermediate individuals in the elliptical pupil parameter population
The larger the contribution of (c), and when CR is 0, the two are equal, thus favoring the diversity and global searching capability of the elliptical pupil parameter population.
Aiming at the relative advantages analyzed by the invention, the cross probability factor with smaller value can be considered in the initial stage of the evolution search, which is beneficial to extremely improving the global search capability of the elliptical pupil parameter, and the cross probability factor with larger value can be used in the later stage of the evolution to accelerate the convergence of the elliptical pupil parameter.
Based on thisThe invention deduces a cross probability factor CR for structure optimization, and assumes that the kernel function of the optimized cross probability factor has incremental property
That is, an approximately upwardly open parabola is used as a kernel function, so that the argument x of the kernel function and the current evolution algebra g of the differential evolution algorithm have a correlation, and thus the argument is made
Then the formula of the optimized cross probability factor
Wherein G is [0, G ]max]G stands for the algebra of the current evolution, GmaxRepresents the maximum number of optimization generations, [ CR ] of the initial settingmin,CRmax]The value range of the cross probability factor is determined based on the application scene. Based on the derivation analysis, it can be understood that the better cross probability factor CR is an increasing function based on the population evolution algebra g of the elliptical pupil parameter, so that the requirement of improving the global search capability of the elliptical pupil location algorithm in the initial stage of the differential evolution can be satisfied, and the requirement of accelerating convergence of the elliptical pupil algorithm in the later stage of the differential evolution can be satisfied.
It should be noted that the foregoing optimization for the two factors in the expansion link may be implemented independently or comprehensively, and those skilled in the art can understand that when the differential evolution ellipse solving process is performed, if the optimized scaling factor and the cross probability factor are comprehensively implemented, the positioning process for the deformed pupil may achieve a better effect.
Finally, the invention also verifies the advantages of the improved differential evolution strategy of the cross probability factor and the scaling factor in the performance of the convergence speed, the convergence precision, the global optimum value search and the like of the positioning of the elliptical pupil through practical tests, compared with the unoptimized differential evolution strategy.
The following referential experiment is based on the pre-collected ideal eye image sample Ea, the eye image sample Ec with the interference factor and the parameter selection of the unoptimized standard differential evolution algorithm as follows: the number of the population is 30, and the maximum iteration number Gmax200, the crossover probability factor CR is 0.9, and the scaling factor F is 0.5; the parameter selection of the improved differential evolution algorithm refers to the following steps: the number of the population is 30, and the maximum iteration number GmaxIs 200, cross probability factor CRmin＝0.2，CRmax0.9, scaling factor Fmin＝0.2，Fmax0.9. Due to the randomness of the differential evolution algorithm, in order to ensure the reliability of the experimental results, 3 rounds of tests were performed on all image samples under the aforementioned experimental parameters, and the mean value of the 3 rounds of test results was taken as the final experimental results, as shown in table 1.
TABLE 1 results of the experiment
As can be seen from table 1, in the present invention, after the coarse positioning processing and the differential evolution strategy are introduced, the positioning processing on the deformed pupil has a high accuracy and a short positioning time, and can solve the problems of low processing efficiency of the conventional deformed pupil positioning scheme, especially the positioning failure caused when the pupil is interfered. The method and the device mainly avoid the pupil edge point extraction step which is relied on by the traditional algorithm, and realize the positioning of the deformed pupil by counting the gray difference values of all the pixel points on the possible ellipse circumference in the area range of the pupil in the input image, thereby showing stronger robustness.
Further, as can be seen from table 1, the optimized differential evolution strategy shows better deformed pupil localization effect, when the standard DE algorithm is used, the evolution algebra is substantially converged when being 145, and particularly, for an ideal eye image, the algebra achieving convergence is less and is 137 times. And even for the eye chart with interference, the differential evolution strategy after the optimization of the individual parameter expansion factors is used, the target parameters can be output only by 85 generations of evolution, the evolution speed is obviously improved, the positioning accuracy is not reduced or increased, and the differential evolution strategy is introduced, particularly the improved differential evolution strategy is applied to the deformed pupil positioning, so that the practical requirements on the pupil positioning under various scenes are completely met.
In summary, the idea of the present invention is to lock the approximate range of the pupil in the eye image by preliminary positioning, and on the basis of this range, aiming at the characteristic that the deformed pupil positioning has higher spatial complexity and time complexity, and combining the pixel gray scale of the eye image and the preset differential evolution strategy, perform global evolutionary parameter search in the preliminary positioning result to accelerate to obtain the optimal solution of the pupil shape parameter. The method can efficiently search the position of the deformed pupil in the input eye moving image, and particularly can obviously improve the speed and the precision of pupil positioning aiming at the condition of the existence of uneven light spots, eyelashes, hair, eyelids, spectacle frames and other interference factors.
Corresponding to the above embodiments and preferred schemes, the present invention further provides an embodiment of a deformed pupil positioning device, as shown in fig. 3, which may specifically include the following components:
the preliminary positioning module 1 is used for carrying out preliminary pupil positioning on the acquired to-be-detected eye movement image to obtain the range of pupil parameters; wherein the pupil parameters include of the pupil: the central abscissa, the central ordinate, the major axis, the major and minor axis proportions and the deflection angle;
the pixel gray scale acquisition module 2 is configured to acquire a gray scale value of a pixel point meeting the pupil parameter from the to-be-detected eye moving image;
and the pupil parameter searching module 3 is used for searching all pixel points in the range according to the gray value and a preset differential evolution strategy and determining the target position of the pupil in the eye image to be detected.
In at least one possible implementation manner, the pupil parameter search module includes:
the pupil parameter solving submodule is used for solving a gray level difference value of all pixel points in the range based on a preset ellipse difference operator;
and the target parameter determining submodule is used for searching an extreme value of the gray level difference value and determining a target pupil parameter according to the difference extreme value.
In at least one possible implementation manner, the pupil parameter search module specifically includes:
the population initialization unit is used for constructing a parameter population based on the range of the pupil parameter, and the parameter population comprises a plurality of parameter individuals;
the parameter individual expanding unit is used for expanding the parameter individuals;
the parameter solving individual unit is used for calculating gray level difference values of all pixel points in the range by combining the parameter individuals before and after expansion and the ellipse difference operator;
the parameter individual selection unit has calculation results of parameter individuals before and after full-bodied comparison expansion, updates the parameter population by using the parameter individual corresponding to the extreme value of the gray level difference value, and evolves to obtain a next generation parameter population;
and the evolution termination unit is used for searching the pupil parameters according to the process until the preset evolution conditions are met and outputting the target pupil parameters.
In at least one possible implementation manner, the parameter individual expansion unit includes:
the individual variation component is used for performing variation on the parameter individuals according to a preset scaling factor to obtain variation individuals corresponding to each parameter individual;
and the individual crossing component is used for generating a plurality of intermediate individuals according to the parameter individuals, the variant individuals and a preset crossing probability factor.
In at least one possible implementation manner, values of scaling factors in the differential evolution strategy are preset to be in a generation-by-generation decreasing trend along with the increase of population evolution generations; and/or the value of the cross probability factor in the differential evolution strategy is preset to be in a generation-by-generation increasing trend along with the increase of population evolution generations.
It should be understood that the division of the various components in the deformed pupil positioning device shown in fig. 3 is merely a logical division, and the actual implementation may be wholly or partially integrated into a physical entity or physically separated. And these components may all be implemented in software invoked by a processing element; or may be implemented entirely in hardware; and part of the components can be realized in the form of calling by the processing element in software, and part of the components can be realized in the form of hardware. For example, a certain module may be a separate processing element, or may be integrated into a certain chip of the electronic device. Other components are implemented similarly. In addition, all or part of the components can be integrated together or can be independently realized. In implementation, each step of the above method or each component above may be implemented by an integrated logic circuit of hardware in a processor element or an instruction in the form of software.
For example, the above components may be one or more integrated circuits configured to implement the above methods, such as: one or more Application Specific Integrated Circuits (ASICs), one or more microprocessors (DSPs), one or more Field Programmable Gate Arrays (FPGAs), etc. For another example, these components may be integrated together and implemented in the form of a System-On-a-Chip (SOC).
In view of the foregoing examples and preferred embodiments thereof, it will be appreciated by those skilled in the art that, in practice, the technical idea underlying the present invention may be applied in a variety of embodiments, the present invention being schematically illustrated by the following vectors:
(1) an electronic device is provided. The device may specifically include: one or more processors, memory, and one or more computer programs, wherein the one or more computer programs are stored in the memory, the one or more computer programs comprising instructions, which when executed by the apparatus, cause the apparatus to perform the steps/functions of the foregoing embodiments or an equivalent implementation.
The electronic device may specifically be a computer-related electronic device, such as but not limited to various interactive terminals and electronic products, and may also be a related electronic device included in the head-mounted gaze tracking system as mentioned above.
Fig. 4 is a schematic structural diagram of an embodiment of an electronic device provided in the present invention, and specifically, the electronic device 900 includes a processor 910 and a memory 930. Wherein, the processor 910 and the memory 930 can communicate with each other and transmit control and/or data signals through the internal connection path, the memory 930 is used for storing computer programs, and the processor 910 is used for calling and running the computer programs from the memory 930. The processor 910 and the memory 930 may be combined into a single processing device, or more generally, separate components, and the processor 910 is configured to execute the program code stored in the memory 930 to implement the functions described above. In particular implementations, the memory 930 may be integrated with the processor 910 or may be separate from the processor 910.
In addition, to further enhance the functionality of the electronic device 900, the device 900 may further include one or more of an input unit 960, a display unit 970, an audio circuit 980, a camera 990, a sensor 901, and the like, which may further include a speaker 982, a microphone 984, and the like. The display unit 970 may include a display screen, among others.
Further, the apparatus 900 may also include a power supply 950 for providing power to various devices or circuits within the apparatus 900.
It should be understood that the operation and/or function of the various components of the apparatus 900 can be referred to in the foregoing description with respect to the method, system, etc., and the detailed description is omitted here as appropriate to avoid repetition.
It should be understood that the processor 910 in the electronic device 900 shown in fig. 4 may be a system on chip SOC, and the processor 910 may include a Central Processing Unit (CPU), and may further include other types of processors, such as: an image Processing Unit (GPU), etc., which will be described in detail later.
In summary, various portions of the processors or processing units within the processor 910 may cooperate to implement the foregoing method flows, and corresponding software programs for the various portions of the processors or processing units may be stored in the memory 930.
(2) A computer data storage medium having stored thereon a computer program or the above apparatus which, when executed, causes a computer to perform the steps/functions of the preceding embodiments or equivalent implementations.
In several embodiments provided by the present invention, any of the functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer data-accessible storage medium. Based on this understanding, some aspects of the present invention may be embodied in the form of software products, which are described below, or portions thereof, which substantially contribute to the art.
In particular, it should be noted that the storage medium may refer to a server or a similar computer device, and specifically, the aforementioned computer program or the aforementioned apparatus is stored in a storage device in the server or the similar computer device.
(3) A computer program product (which may include the above apparatus), when run on a terminal device, causes the terminal device to perform the deformed pupil positioning method of the foregoing embodiments or equivalent implementations.
From the above description of the embodiments, it is clear to those skilled in the art that all or part of the steps in the above implementation method can be implemented by software plus a necessary general hardware platform. With this understanding, the above-described computer program product may include, but is not limited to referring to APP.
In the foregoing, the device/terminal may be a computer device, and the hardware structure of the computer device may further specifically include: at least one processor, at least one communication interface, at least one memory, and at least one communication bus; the processor, the communication interface and the memory can all complete mutual communication through the communication bus. The processor may be a central Processing unit CPU, a DSP, a microcontroller, or a digital Signal processor, and may further include a GPU, an embedded Neural Network Processor (NPU), and an Image Signal Processing (ISP), and may further include a specific integrated circuit ASIC, or one or more integrated circuits configured to implement the embodiments of the present invention, and the processor may have a function of operating one or more software programs, and the software programs may be stored in a storage medium such as a memory; and the aforementioned memory/storage media may comprise: non-volatile memories (non-volatile memories) such as non-removable magnetic disks, U-disks, removable hard disks, optical disks, etc., and Read-Only memories (ROM), Random Access Memories (RAM), etc.
In the embodiments of the present invention, "at least one" means one or more, "a plurality" means two or more. "and/or" describes the association relationship of the associated objects, and means that there may be three relationships, for example, a and/or B, and may mean that a exists alone, a and B exist simultaneously, and B exists alone. Wherein A and B can be singular or plural. The character "/" generally indicates that the former and latter associated objects are in an "or" relationship. "at least one of the following" and similar expressions refer to any combination of these items, including any combination of singular or plural items. For example, at least one of a, b, and c may represent: a, b, c, a and b, a and c, b and c or a and b and c, wherein a, b and c can be single or multiple.
Those of skill in the art will appreciate that the various modules, elements, and method steps described in the embodiments disclosed in this specification can be implemented as electronic hardware, combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
And, modules, units, etc. described herein as separate components may or may not be physically separate, i.e., may be located in one place, or may be distributed across multiple places, e.g., nodes of a system network. Some or all of the modules and units can be selected according to actual needs to achieve the purpose of the above-mentioned embodiment. Can be understood and carried out by those skilled in the art without inventive effort.
The structure, features and effects of the present invention have been described in detail with reference to the embodiments shown in the drawings, but the above embodiments are merely preferred embodiments of the present invention, and it should be understood that technical features related to the above embodiments and preferred modes thereof can be reasonably combined and configured into various equivalent schemes by those skilled in the art without departing from and changing the design idea and technical effects of the present invention; therefore, the invention is not limited to the embodiments shown in the drawings, and all the modifications and equivalent embodiments that can be made according to the idea of the invention are within the scope of the invention as long as they are not beyond the spirit of the description and the drawings.
Claims (10)
1. A deformed pupil positioning method is characterized by comprising the following steps:
performing pupil preliminary positioning on the obtained to-be-detected eye moving image to obtain the range of pupil parameters; wherein the pupil parameters include of the pupil: the central abscissa, the central ordinate, the major axis, the major and minor axis proportions and the deflection angle;
acquiring the gray value of the pixel point meeting the pupil parameter in the eye moving image to be detected;
and searching all pixel points which accord with the range according to the gray value and a preset differential evolution strategy and determining the target position of the pupil in the eye image to be detected.
2. The method for positioning deformed pupils according to claim 1, wherein the searching all the pixels that meet the range according to the gray-level value and a preset differential evolution strategy comprises:
based on a preset ellipse difference operator, solving gray level difference values of all pixel points in the range;
searching an extreme value of the gray level difference value, and determining a target pupil parameter according to the difference extreme value.
3. The method for positioning deformed pupils according to claim 2, wherein the searching for all the pixels that meet the range according to the gray-level value and a preset differential evolution strategy specifically comprises:
constructing a parameter population based on the range of the pupil parameter, wherein the parameter population comprises a plurality of parameter individuals;
expanding the parameter individuals;
calculating gray level difference values of all pixel points in the range by combining the parameter individuals before and after expansion and the ellipse difference operator;
comparing the calculation results of the parameter individuals before and after expansion, updating the parameter population by using the parameter individuals corresponding to the extreme value of the gray level difference value, and evolving to obtain a next generation parameter population;
and searching the pupil parameters according to the process until the preset evolution condition is met, and outputting the target pupil parameters.
4. The deformed pupil positioning method according to claim 3, wherein expanding the individual according to the parameters comprises:
carrying out variation on the parameter individuals according to a preset scaling factor to obtain variation individuals corresponding to each parameter individual;
and generating a plurality of intermediate individuals according to the parameter individuals, the variant individuals and a preset cross probability factor.
5. The deformed pupil positioning method according to claim 4, wherein the scaling factor has a decreasing value with increasing population evolution generations, and/or the cross probability factor has an increasing value with increasing population evolution generations.
6. A deformed pupil positioning device, comprising:
the preliminary positioning module is used for carrying out preliminary pupil positioning on the acquired to-be-detected eye movement image to obtain the range of pupil parameters; wherein the pupil parameters include of the pupil: the central abscissa, the central ordinate, the major axis, the major and minor axis proportions and the deflection angle;
the pixel gray level acquisition module is used for acquiring the gray level value of the pixel point meeting the pupil parameter in the eye moving image to be detected;
and the pupil parameter searching module is used for searching all pixel points in the range according to the gray value and a preset differential evolution strategy and determining the target position of the pupil in the eye image to be detected.
7. The deformed pupil positioning device of claim 6, wherein the pupil parameter search module comprises:
the pupil parameter solving submodule is used for solving a gray level difference value of all pixel points in the range based on a preset ellipse difference operator;
and the target parameter determining submodule is used for searching an extreme value of the gray level difference value and determining a target pupil parameter according to the difference extreme value.
8. The pupil deformation positioning device according to claim 6 or 7, wherein the values of the scaling factors in the differential evolution strategy are preset to be in a generation-by-generation decreasing trend with the increase of population evolution generations; and/or the value of the cross probability factor in the differential evolution strategy is preset to be in a generation-by-generation increasing trend along with the increase of population evolution generations.
9. An electronic device, comprising:
one or more processors, a memory, and one or more computer programs, wherein the one or more computer programs are stored in the memory, the one or more computer programs comprising instructions which, when executed by the apparatus, cause the apparatus to perform the deformed pupil positioning method of any of claims 1-5.
10. A computer data storage medium having a computer program stored therein, which when run on a computer causes the computer to perform the deformed pupil positioning method of any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110369394.8A CN112926536B (en) | 2021-04-06 | 2021-04-06 | Deformed pupil positioning method, device and equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110369394.8A CN112926536B (en) | 2021-04-06 | 2021-04-06 | Deformed pupil positioning method, device and equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112926536A true CN112926536A (en) | 2021-06-08 |
| CN112926536B CN112926536B (en) | 2024-04-16 |
Family
ID=76174304
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110369394.8A Active CN112926536B (en) | 2021-04-06 | 2021-04-06 | Deformed pupil positioning method, device and equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112926536B (en) |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101266645A (en) * | 2008-01-24 | 2008-09-17 | 电子科技大学中山学院 | Iris positioning method based on multi-resolutions analysis |
| CN102800107A (en) * | 2012-07-06 | 2012-11-28 | 浙江工业大学 | Motion target detection method based on improved minimum cross entropy |
| CN103164704A (en) * | 2013-04-12 | 2013-06-19 | 山东师范大学 | Iris image segmentation algorithm based on mixed Gaussian model |
| US20160045109A1 (en) * | 2014-08-14 | 2016-02-18 | Utechzone Co., Ltd. | Method, apparatus and computer program product for positioning pupil |
| CN106056092A (en) * | 2016-06-08 | 2016-10-26 | 华南理工大学 | Gaze estimation method for head-mounted device based on iris and pupil |
| CN108256392A (en) * | 2016-12-29 | 2018-07-06 | 广州映博智能科技有限公司 | Pupil region localization method based on projecting integral and area grayscale extreme value |
| CN109646198A (en) * | 2019-02-19 | 2019-04-19 | 西安科技大学 | A kind of electric wheelchair control system and method for view-based access control model tracking |
| CN109857254A (en) * | 2019-01-31 | 2019-06-07 | 京东方科技集团股份有限公司 | Pupil positioning method and device, VR/AR equipment and computer-readable medium |
| US20190347824A1 (en) * | 2018-05-14 | 2019-11-14 | Beijing Boe Optoelectronics Technology Co., Ltd. | Method and apparatus for positioning pupil, storage medium, electronic device |
| CN111339982A (en) * | 2020-03-05 | 2020-06-26 | 西北工业大学 | Multi-stage pupil center positioning technology implementation method based on features |
| US20200305705A1 (en) * | 2018-11-12 | 2020-10-01 | Wenzhou Medical University | Method for processing pupil tracking image |
| CN111854620A (en) * | 2020-07-16 | 2020-10-30 | 科大讯飞股份有限公司 | Monocular camera-based actual pupil distance measuring method, device and equipment |
| CN112434675A (en) * | 2021-01-26 | 2021-03-02 | 西南石油大学 | Pupil positioning method for global self-adaptive optimization parameters |
-
2021
- 2021-04-06 CN CN202110369394.8A patent/CN112926536B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101266645A (en) * | 2008-01-24 | 2008-09-17 | 电子科技大学中山学院 | Iris positioning method based on multi-resolutions analysis |
| CN102800107A (en) * | 2012-07-06 | 2012-11-28 | 浙江工业大学 | Motion target detection method based on improved minimum cross entropy |
| CN103164704A (en) * | 2013-04-12 | 2013-06-19 | 山东师范大学 | Iris image segmentation algorithm based on mixed Gaussian model |
| US20160045109A1 (en) * | 2014-08-14 | 2016-02-18 | Utechzone Co., Ltd. | Method, apparatus and computer program product for positioning pupil |
| CN106056092A (en) * | 2016-06-08 | 2016-10-26 | 华南理工大学 | Gaze estimation method for head-mounted device based on iris and pupil |
| CN108256392A (en) * | 2016-12-29 | 2018-07-06 | 广州映博智能科技有限公司 | Pupil region localization method based on projecting integral and area grayscale extreme value |
| US20190347824A1 (en) * | 2018-05-14 | 2019-11-14 | Beijing Boe Optoelectronics Technology Co., Ltd. | Method and apparatus for positioning pupil, storage medium, electronic device |
| US20200305705A1 (en) * | 2018-11-12 | 2020-10-01 | Wenzhou Medical University | Method for processing pupil tracking image |
| CN109857254A (en) * | 2019-01-31 | 2019-06-07 | 京东方科技集团股份有限公司 | Pupil positioning method and device, VR/AR equipment and computer-readable medium |
| CN109646198A (en) * | 2019-02-19 | 2019-04-19 | 西安科技大学 | A kind of electric wheelchair control system and method for view-based access control model tracking |
| CN111339982A (en) * | 2020-03-05 | 2020-06-26 | 西北工业大学 | Multi-stage pupil center positioning technology implementation method based on features |
| CN111854620A (en) * | 2020-07-16 | 2020-10-30 | 科大讯飞股份有限公司 | Monocular camera-based actual pupil distance measuring method, device and equipment |
| CN112434675A (en) * | 2021-01-26 | 2021-03-02 | 西南石油大学 | Pupil positioning method for global self-adaptive optimization parameters |
Non-Patent Citations (5)
| Title |
|---|
| HONGPING DENG ET.AL: "A Robust Algorithm for Distorted Pupil Localization based on Ellipse Difference and PSO", 《2010 3RD INTERNATIONAL CONGRESS ON IMAGE AND SIGNAL PROCESSING (CISP2010)》, pages 2478 - 2482 * |
| ZOU DE-XUAN ET.AL: "Iris location algorithm based on modified differential evolution algorithm", 《 CONTROL THEORY & APPLICATIONS》, pages 1194 - 1200 * |
| 高岳林 等: "一种带有随机变异的动态差分进化算法", 《计算机应用》, vol. 29, no. 10, pages 2719 - 2722 * |
| 高源;梁正友;: "基于融合星射线法和椭圆拟合法的瞳孔定位研究", 电子世界, no. 07, pages 147 * |
| 龚鼎盛 等: "基于抛物线拟合和边缘点搜索的虹膜精确定位", 《微型机与应用》, pages 45 - 47 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112926536B (en) | 2024-04-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Lezama et al. | Finding vanishing points via point alignments in image primal and dual domains | |
| CN107481279B (en) | Monocular video depth map calculation method | |
| US9280821B1 (en) | 3-D reconstruction and registration | |
| Ma et al. | An image matching optimization algorithm based on pixel shift clustering RANSAC | |
| CN111767960A (en) | Image matching method and system applied to image three-dimensional reconstruction | |
| CN110490913A (en) | Feature based on angle point and the marshalling of single line section describes operator and carries out image matching method | |
| CN110930411A (en) | Human body segmentation method and system based on depth camera | |
| Li et al. | Coarse-to-fine PatchMatch for dense correspondence | |
| CN114782499A (en) | Image static area extraction method and device based on optical flow and view geometric constraint | |
| Shi et al. | Robust image registration using structure features | |
| Lecca et al. | Comprehensive evaluation of image enhancement for unsupervised image description and matching | |
| CN106295710A (en) | Image local feature matching process, device and terminal of based on non-geometric constraint | |
| Liu et al. | Unsupervised global and local homography estimation with motion basis learning | |
| CN110728296B (en) | Two-step random sampling consistency method and system for accelerating feature point matching | |
| Chen et al. | Multi-stage matching approach for mobile platform visual imagery | |
| Lin et al. | Matching cost filtering for dense stereo correspondence | |
| CN112926536A (en) | Deformed pupil positioning method, device and equipment | |
| Schnabel et al. | Shape detection in point clouds | |
| Kim et al. | Real-time stereo matching using extended binary weighted aggregation | |
| CN115082784A (en) | Multi-temporal remote sensing image unsupervised change detection pseudo sample automatic generation method | |
| Jiang et al. | Progressive filtering for feature matching | |
| Liu et al. | A new ORB feature matching image algorithm based on non-maximum suppression method and Retinal sampling model | |
| Tang et al. | A GMS-guided approach for 2D feature correspondence selection | |
| Gonzalez-Huitron et al. | Jaccard distance as similarity measure for disparity map estimation | |
| Sun et al. | Binocular Ranging Based on Improved ORB-RANSAC |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |