A method for segmenting region of interest in finger vein images based on active contour method

Through the active contour method of the region of interest segmentation method of finger vein image, the finger edge detection and iterative processing is performed using the Sobel operator and Chan-Vese model, combined with morphological operation and connectivity domain analysis, the ROI segmentation effect in the prior art is solved Poor problems, achieving more robust and efficient finger vein image segmentation.

CN114241186BInactive Publication Date: 2024-11-08CHANGCHUN UNIV OF TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202111403561.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-11-24
Publication Date
2024-11-08
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The existing ROI segmentation method cannot obtain a robust segmentation effect when processing finger vein images, especially images acquired for different acquisition devices. It has a large amount of calculation and is not efficient. The gradient operator-based method is at the edge. When it is not obvious, it is easy to mistakenly think that the background noise is the edge, resulting in incorrect segmentation.

Method used

The region of interest segmentation method of finger vein image based on the active contour method is used to detect finger edges through horizontal Sobel operators, and iterative processing is performed in combination with the Chan-Vese model to remove the small-area connective domain, and combine morphological operations and connective domain analysis. , realize the segmentation of finger vein images.

Benefits of technology

It realizes more robust ROI segmentation, which can effectively remove background noise, improve the continuity and integrity of segmentation, improve the matching performance of finger vein images, and improve the efficiency and accuracy of the system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114241186B_ABST
    Figure CN114241186B_ABST
Patent Text Reader

Abstract

The present invention proposes a method for segmenting the region of interest of a finger vein image based on an active contour method, comprising the following steps: Step 1, using a horizontal Sobel operator to detect the edge of the finger, obtaining a series of maximum response value points as an initial horizontal point set for marking the edge of the finger; Step 2, iteratively processing the initial horizontal point set through a Chan-Vese model to obtain a finger edge image containing a small area connected domain; Step 3, through post-processing operations and combined with connected domain analysis, removing the connected domain smaller than 1 / 10 of the image size, obtaining a finger vein image segmentation mask, and completing the segmentation of the region of interest. The method of the present invention is a process of expanding from local to global, with better robustness, and combined with some morphological operations and connected domain analysis processing, a very ideal effect is obtained.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of image processing and biometric identification, and in particular to a method for segmenting a region of interest (ROI) of a finger vein image based on an active contour method. Background Art

[0002] Finger vein recognition is a recognition technology based on biological physiological characteristics. The veins in the finger are captured by near-infrared light (700nm-1000nm) to produce shadows, thus forming a vein image. Compared with other biological physiological characteristics, finger vein features have very obvious advantages: (1) Liveness recognition: Finger vein images can only be obtained on living humans; (2) Uniqueness: Each person's finger vein pattern is unique; (3) Stability: Each person's finger vein distribution characteristics remain unchanged throughout life after adulthood; (4) Security: Finger veins are distributed under the skin, and the pattern is complex and difficult to forge. Therefore, in recent years, finger vein recognition has important application value in the fields of identity authentication, access management, security monitoring, e-commerce, e-government, etc. due to its series of advantages, and has become one of the important research fields in biometric recognition technology.

[0003] Generally, finger vein recognition is divided into four steps: image acquisition, preprocessing, feature extraction, and matching authentication. In the process of finger vein image acquisition, due to the influence of factors such as finger displacement, rotation, complex background, noise and lighting, the vein images collected from the same finger are often different. Therefore, no matter which feature extraction method is used, good feature extraction cannot be performed, resulting in a decrease in matching performance. Therefore, the preprocessing process is very important for the finger vein recognition system. Usually, preprocessing can be divided into region of interest segmentation, vein image enhancement, and image normalization. The accurate segmentation of ROI is the premise of vein image enhancement and image normalization, and ROI segmentation is also a method that can remove background noise to the greatest extent.

[0004] Compared with the existing ROI segmentation methods, the threshold-based ROI segmentation method often cannot obtain a robust segmentation effect when processing images acquired by different acquisition devices. The image edges obtained when binarizing the image are not smooth and continuous. For some poor-quality finger vein images, even the correct ROI image cannot be obtained. In addition, this method has a large amount of calculation and is inefficient. The ROI segmentation method based on the gradient operator cannot obtain continuous edges when processing images with unclear or even disappeared finger edges. When looking for the maximum response value of the upper and lower parts of the horizontal gradient image and using it as the finger edge, the background noise will be mistaken for the edge, resulting in incorrect segmentation of the ROI area.

[0005] By observing the finger vein image, it is found that in general captured images, the background area is usually darker (smaller pixel value), the finger area is brighter (larger pixel value), and because the finger edge is thinner, more light can pass through the inner area of ​​the finger edge, so it is the brightest part in the image. Based on this principle, the present invention provides a finger area segmentation algorithm based on the active contour method. Compared with the threshold method and the gradient operator method, it is a process of expanding from local to global, has better robustness, and combined with some post-processing operations, a very ideal effect is obtained. Summary of the invention

[0006] In order to solve the above problems, it is necessary to provide a method for segmenting the region of interest of finger vein images based on the active contour method.

[0007] A first aspect of the present invention provides a method for segmenting a region of interest of a finger vein image based on an active contour method, comprising the following steps:

[0008] Step 1: Use the horizontal Sobel operator to detect the finger edge and obtain a series of maximum response value points as the initial horizontal point set for marking the finger edge;

[0009] Step 2, iteratively process the initial horizontal point set through the Chan-Vese model to obtain a finger edge image containing a small area connected domain;

[0010] Step 3: Through post-processing operations and combined with connected domain analysis, the connected domains smaller than 1 / 10 of the image size are removed to obtain the finger vein image segmentation mask and complete the segmentation of the ROI.

[0011] Based on the above, the Chan-Vese model includes:

[0012] The initial horizontal point set obtained is recorded as

[0013] Through the idea of ​​variational method, the initial level set is obtained Iterate formula (1) to find the energy function in the Chan-Vese model Minimum value of:

[0014]

[0015] in, is the level set function The functional of I(x) is the two-dimensional plane curve to be solved, λ 1 represents the constant parameter, c 1 ,c 2 represents a known two-dimensional plane curve, represents a step function;

[0016] for Adding a small perturbation term v(x), we get:

[0017]

[0018] express The integrand in ;

[0019] When the functional When it is a minimum value, it satisfies:

[0020]

[0021] In guarantee and When the direction is opposite, it is achieved:

[0022]

[0023] get Gradient descent flow for the expression:

[0024]

[0025] Combined with the iterative expression of formula (1), the level set get:

[0026]

[0027] The stable solution of the gradient descent flow is used as the solution of the Euler-Lagrange equation corresponding to the variational problem, and the iterative process is performed:

[0028]

[0029] A finger edge image containing a small connected domain can be obtained.

[0030] Based on the above, step 3 specifically includes:

[0031] Step 3.1, performing dilation and erosion operations on the finger edge image obtained in step 2;

[0032] Step 3.2, for the image after dilation and erosion operations, combined with connected domain analysis, remove the connected domains in the image that are smaller than 1 / 10 of the image size;

[0033] In step 3.3, the upper and lower finger edges obtained in step 3.2 are added to the upper and lower finger edges detected using the Sobel edge detection operator and the median is taken to obtain a complete finger edge processing result, that is, a finger vein image segmentation mask.

[0034] A second aspect of the present invention provides a terminal, comprising a processor, a memory, and a finger vein image preprocessing algorithm program stored in the memory. When the finger vein image preprocessing algorithm program is executed by the processor, the steps of the finger vein image region of interest segmentation algorithm based on the active contour method are implemented.

[0035] A third aspect of the present invention provides a computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the steps of the algorithm for segmenting a region of interest of a finger vein image based on the active contour method.

[0036] Compared with the prior art, the present invention has outstanding substantive features and significant progress. Specifically, the method of the present invention first performs Gaussian filtering on the obtained image to remove salt and pepper noise; secondly, an edge tracking starting point is set, and the Sobel operator is selected to detect the finger edge. The maximum response value of the Sobel operator is considered to be the finger edge, and the initial level set is set on this edge so that the Chan-Vese model used can accelerate its convergence speed; then the Chan-Vese model is used for iteration to obtain the initial level set contour; finally, through partial morphological operations and connected domain analysis, the final finger segmentation mask is obtained to achieve the segmentation of the region of interest of the finger vein image. The method of the present invention is a process of expanding from local to global, with better robustness, and combined with partial morphological operations and connected domain analysis processing, a very ideal effect is obtained.

[0037] Additional aspects and advantages of the present invention will become apparent from the following description or may be learned by practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is a flow chart of the method of the present invention.

[0039] Figure 2 It is a schematic diagram of the principle of the method of the present invention.

[0040] Figure 3 It is a comparison effect diagram of the method of the present invention, the threshold method and the mask method. DETAILED DESCRIPTION

[0041] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0042] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0043] Example 1

[0044] like Figure 1 As shown, this embodiment provides a method for segmenting a region of interest of a finger vein image based on an active contour method, comprising the following steps:

[0045] Step 1: Use the horizontal Sobel operator to detect the finger edge and obtain a series of maximum response value points as the initial horizontal point set for marking the finger edge;

[0046] Step 1.1, perform Gaussian filtering on the collected finger vein image, and the resulting image is denoted as F;

[0047] Step 1.2, for the image F obtained in step 1.1, use the horizontal Sobel operator to perform a convolution operation on the entire image to obtain the horizontal gradient map of the finger vein image, which is denoted as G; select a column of the horizontal gradient image G as the baseline, and find the maximum response values ​​of its upper and lower parts. Similarly, a series of horizontal points with the maximum response values ​​of the Sobel operator are denoted as the initial level set.

[0048] Step 2, iteratively process the initial horizontal point set through the Chan-Vese model to obtain a finger edge image containing a small area connected domain;

[0049] Specifically, Figure 2 As shown, the Chan-Vese model includes:

[0050] The initial horizontal point set obtained is recorded as As shown in Figure (2.a);

[0051] The initial level set obtained by the variational method Iterate formula (1) to find the energy function in the Chan-Vese model The minimum value of , the search process is shown in Figure (2.b):

[0052]

[0053] in, is the level set function The functional of I(x) is the two-dimensional plane curve to be solved, λ 1 represents the constant parameter, c 1 ,c 2 represents a known two-dimensional plane curve, represents a step function;

[0054] for Adding a small perturbation term v(x), we get:

[0055]

[0056] express The integrand in ;

[0057] When the functional When it is a minimum value, it satisfies:

[0058]

[0059] In guarantee and When the direction is opposite, it is achieved:

[0060]

[0061] get Gradient descent flow for the expression:

[0062]

[0063] Combined with the iterative expression of formula (1), the level set get:

[0064]

[0065] The stable solution of the gradient descent flow is used as the solution of the Euler-Lagrange equation corresponding to the variational problem, and the iterative process is performed:

[0066]

[0067] Then we can get the finger edge image containing a small connected domain, such as Figure 2 As shown in (c) in .

[0068] The Chan-Vese model based on level set is a kind of active contour method. The active contour method defines an "energy" functional (generally composed of internal energy terms and external energy terms) with a continuous curve as an independent variable, and determines the target boundary by finding the curve that minimizes the functional, thereby converting the image segmentation process into an intuitive function solving process. In this embodiment, since it is necessary to find the function corresponding to the minimum value of the functional in the Chan-Vese model, the solution is obtained by the variational method.

[0069] Step 3, through post-processing operations and combined with connected domain analysis, the connected domains smaller than 1 / 10 of the image size are removed to obtain the finger vein image segmentation mask and complete the segmentation of the region of interest ROI;

[0070] Step 3.1, performing dilation and erosion operations on the finger edge image obtained in step 2;

[0071] Step 3.2, for the image after dilation and erosion operations, combined with connected domain analysis, remove the connected domains in the image that are smaller than 1 / 10 of the image size;

[0072] Step 3.3, add the upper and lower finger edges obtained in step 3.2 to the upper and lower finger edges detected using the Sobel edge detection operator and take the median to obtain the complete finger edge processing result, that is, the finger vein image segmentation mask, as shown in Figure 2 As shown in (d) in .

[0073] Figure 3 The comparison between the method of this embodiment and the threshold method and mask method is provided to highlight the advantages of the method in this embodiment. The first column is the original finger vein image collected, the second column is the finger mask segmentation result obtained based on the method in this embodiment, the third column is the finger mask segmentation result obtained based on the threshold method, and the fourth column is the finger mask segmentation result obtained based on the gradient operator method. Figure 3 It can be seen from the segmentation comparison results in that by comparing the threshold method with the gradient operator method, the finger vein image region of interest segmentation method based on the active contour method mentioned in this embodiment can extract the complete finger edge and retain the finger vein region of interest to the greatest extent.

[0074] Example 2

[0075] This embodiment provides a terminal, including a processor, a memory, and a finger vein image preprocessing algorithm program stored in the memory. When the finger vein image preprocessing algorithm program is executed by the processor, the steps of the finger vein image region of interest segmentation algorithm based on the active contour method as described in Example 1 are implemented.

[0076] Example 3

[0077] This embodiment provides a computer-readable storage medium on which computer instructions are stored. When the computer instructions are executed by a processor, the steps of the algorithm for segmenting the region of interest of a finger vein image based on the active contour method as described in Embodiment 1 are implemented.

[0078] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0079] Each functional unit in each embodiment of the present application 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.

[0080] If the above-mentioned integrated module 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 present application implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The above-mentioned computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the above-mentioned computer program includes computer program code, and the above-mentioned computer program code can be in source code form, object code form, executable file or some intermediate form, etc.

[0081] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for segmenting a region of interest in a finger vein image based on an active contour method, characterized in that: The following steps are involved: Step 1: Use the horizontal Sobel operator to detect the finger edge and obtain a series of maximum response value points as the initial horizontal point set for marking the finger edge; Step 1.1, perform Gaussian filtering on the collected finger vein image, and the resulting image is recorded as F ; Step 1.2: for the image obtained in step 1.1 F , the whole image is convolved using the horizontal Sobel operator, and the horizontal gradient map of the finger vein image is obtained as G ; Select horizontal gradient image G A column is used as the baseline, and the maximum response values ​​of the upper and lower parts are found. Similarly, a series of level points with the maximum response value for the Sobel operator are recorded as the initial level set. Step 2, iteratively process the initial horizontal point set through the Chan-Vese model to obtain a finger edge image containing a small area connected domain; The Chan-Vese model includes: The initial horizontal point set obtained is recorded as ; The initial level set obtained by the variational method Iterate formula (1) to find the energy function in the Chan-Vese model Minimum value of: (1) in, is the level set function The functional of represents the two-dimensional plane curve to be found, represents the constant term parameter, represents a known two-dimensional plane curve, represents a step function; for Adding a small perturbation ,have to: express The integrand in ; When the functional When it is a minimum value, it satisfies: In guarantee and When the direction is opposite, it is achieved: get Gradient descent flow for the expression: Combined with the iterative expression of formula (1), the level set get: The stable solution of the gradient descent flow is used as the solution of the Euler-Lagrange equation corresponding to the variational problem, and the iterative process is performed: The finger edge image containing a small area connected domain can be obtained; Step 3: Through post-processing operations and combined with connected domain analysis, the connected domains smaller than 1 / 10 of the image size are removed to obtain the finger vein image segmentation mask and complete the ROI segmentation; Step 3 specifically includes: Step 3.1, performing dilation and erosion operations on the finger edge image obtained in step 2; Step 3.2, for the image after dilation and erosion operations, combined with connected domain analysis, remove the connected domains in the image that are smaller than 1 / 10 of the image size; In step 3.3, the upper and lower finger edges obtained in step 3.2 are added to the upper and lower finger edges detected using the Sobel edge detection operator and the median is taken to obtain a complete finger edge processing result, that is, a finger vein image segmentation mask.

2. A terminal, characterized in that: The invention comprises a processor, a memory and a finger vein image preprocessing algorithm program stored in the memory. When the finger vein image preprocessing algorithm program is executed by the processor, the steps of the finger vein image region of interest segmentation algorithm based on the active contour method as claimed in claim 1 are implemented.

3. A computer-readable storage medium having computer instructions stored thereon, characterized in that: When the computer instructions are executed by the processor, the steps of the algorithm for segmenting the region of interest of a finger vein image based on the active contour method as claimed in claim 1 are implemented.