Improved directional filter and flexible matching based recognition method

A technology of directional filtering and recognition methods, which is applied in the directions of character and pattern recognition, instruments, computer parts, etc., can solve the problems of great influence on the accuracy of recognition results, large amount of calculation, and great influence on image quality recognition results, etc., to achieve The effect of high reliability, high precision and high use value

Active Publication Date: 2014-10-29
华正清圆(天津)科技有限公司
4 Cites 9 Cited by

AI-Extracted Technical Summary

Problems solved by technology

[0005] Miura et al. proposed a method based on repeated line tracing for vein extraction. The vein extraction effect of this method is better, but the image quality has a greater impact on the recognition results.
Badawi et al. used a point-by-point pixel comparison method for vein matching on the preprocessed vein images. This ...
View more

Abstract

The invention relates to an improved directional filter and flexible matching based recognition method. The gray level difference between finger vein image blood vessels collected under the near-infrared light conditions and the background is small, the contrast ratio is low, and accordingly the subsequent matching recognition of a finger vein image is affected. The improved directional filter and flexible matching based recognition method comprises designing out the neighborhood direction which is suitable for every pixel point of the read finger vein image through the directional characteristics of vein patterns to form a direction template; solving a directional pattern of the vein image through the template; performing direction filtering on the image according to the direction of the pixel points; extracting detail feature points of the image on the basis of the direction filtering; implementing the matching recognition between the feature points within the polar angle range and the polar radius range through the flexible matching thought, wherein the polar angle range is selected to be 1 to 4 degrees and the polar radius range is selected to be 1 to 4 pixel points. The improved directional filter and flexible matching based recognition method is applied to the mode recognition technical field.

Application Domain

Character and pattern recognition

Technology Topic

Near infrared lightGray level +3

Image

  • Improved directional filter and flexible matching based recognition method
  • Improved directional filter and flexible matching based recognition method
  • Improved directional filter and flexible matching based recognition method

Examples

  • Experimental program(6)

Example Embodiment

[0036] Example 1:
[0037] A recognition method based on improved direction filtering and flexible matching. First, use the direction characteristics of the vein pattern to design a direction template suitable for each pixel of the finger vein image using the direction characteristics of the finger vein image, and then filter the direction. The template design expands the traditional 7×7 template into a 9×9 template to solve the template overflow problem that occurs when the template is rotated in the horizontal direction to obtain the template in other directions; on this basis, the idea of ​​flexible matching is used , Realize the matching and recognition between the detailed feature point sets within a certain angle and radius. Such as figure 1 Shown
[0038] From the topological structure of finger veins, it can be known that the vein pattern extends in a specific direction and has obvious directionality. However, the traditional filter does not take the direction information of the vein into consideration, so its filtering enhancement effect is not ideal. The present invention fully considers this characteristic of vein pattern and adopts directional filtering and enhancement processing for vein image, namely obtaining the neighborhood direction of each pixel in the vein image to form a directional map and designing filter templates in multiple directions to form a directional filter When filtering, select the filter in the corresponding direction according to the direction of the pixel to filter the image.

Example Embodiment

[0039] Implementation case 2:
[0040] According to the recognition method based on improved direction filtering and flexible matching described in embodiment 1, the neighborhood direction filtering template of pixel points of the finger vein image is set to 8 neighborhoods of each pixel point. With this pixel as the center of the discrete direction, the reference pixel is located at the center of the direction filter template. The values ​​1-8 on the direction filter template respectively correspond to 8 discrete directions, that is, starting from the horizontal position, press Counterclockwise, determine a direction every 22.5°, the direction angle range of the direction filter template is [0, п) and the template is located in the five adjacent directions of 1, 2, 3, 7, and 8 in the horizontal direction. There are more pixels in the domain, and fewer neighborhood pixels are taken in the vertical directions of 4, 5, and 6, so the directional filter template design is more in line with the finger vein pattern extending roughly to both ends of the finger. Features of gentle direction change. Such as figure 2 Shown
[0041] The determination of the specific direction of the pixel: the grayscale difference between the pixel point and the point on the grain in the same direction is the smallest, and the grayscale difference in the vertical direction is the largest. According to this principle, the above-mentioned neighborhood template is used to find each pixel. Pixel gray average value in 8 directions centered on this point And will Divide them into 4 groups in a perpendicular direction, and calculate the absolute value of the difference between the two averages in each group. , which is
[0042] (1)
[0043] In the above formula, Direction of vein pattern.
[0044] Which makes Two directions in the maximum direction group with +4 for pixels Possible thread direction. take with +4 average gray scale and The direction in which the gray value is close is Line direction ,which is:
[0045] (2)
[0046] As mentioned above, the direction map of the vein image can be obtained by finding the direction of each pixel. In order to reduce the influence of noise and ensure the accuracy of the direction field estimation, an 8×8 sliding window is used to smooth the point pattern to obtain a continuous smooth pattern, such as image 3 Shown.
[0047] The directional filter is a series of filters related to the direction of the pixel. The present invention draws on the principle of O’Gorman et al. [13] to design the directional filter to design the filter coefficients. Since the direction of finger vein pattern mainly extends to both ends, the coefficient of the filter is modified here to make it attenuate from the center to both ends. At the same time, to solve the problem of template overflow caused by rotating the horizontal template to obtain templates in other directions, Will traditional The template expands to the size Template, horizontal filter template, such as Figure 4 As shown, the new expansion coefficients of the outermost circle of the template are all 0. When using this template to obtain templates in other directions, only the center of the template Coordinate rotation is performed on the coefficients in the range, and the position without coefficient in the rotated template is represented by 0, such as Figure 5 As shown, this avoids the situation of discarding coefficients when the template rotates and overflows, and does not destroy the coefficient distribution law.
[0048] After obtaining the eight directional filtering templates, the directional filtering operation is performed on the vein image. During filtering, the directional characteristics of a certain block area are obtained according to the directional map, and a filter in the corresponding direction is selected from a series of filters to filter this block, thereby obtaining an image enhanced by directional filtering, such as Image 6 As shown,

Example Embodiment

[0049] Implementation case 3:
[0050] According to the recognition method based on improved directional filtering and flexible matching described in embodiment 1 or 2, the extended directional filtering template aims at the problem of template overflow that occurs when the template is rotated in the horizontal direction to obtain other direction templates. Extend the traditional 7×7 template to a size of 9×9. The coefficients of the outermost circle of the expanded template are all 0. When using this template to obtain templates in other directions, only the 7×7 center of the template Coordinate rotation is performed on the coefficients in the range, and the position without coefficients in the rotated template is represented by 0. This avoids the situation of discarding coefficients when the template rotates and overflows, and does not destroy the coefficient distribution law.

PUM

no PUM

Description & Claims & Application Information

We can also present the details of the Description, Claims and Application information to help users get a comprehensive understanding of the technical details of the patent, such as background art, summary of invention, brief description of drawings, description of embodiments, and other original content. On the other hand, users can also determine the specific scope of protection of the technology through the list of claims; as well as understand the changes in the life cycle of the technology with the presentation of the patent timeline. Login to view more.

Similar technology patents

Classification method for waste concrete block treatment

InactiveCN111847949AImprove use valueReduce mining of stone
Owner:汉江城建集团有限公司

Numerical control system industrial data collection and storage method and system

ActiveCN110858125AImprove use valueReduce chances of data loss
Owner:HUAZHONG UNIV OF SCI & TECH

Method for producing calcium lactate by taking duck egg shell as raw material

InactiveCN102690191AImprove use valuesolve the pollution problem
Owner:GUANGXI UNIVERSITY OF TECHNOLOGY

Classification and recommendation of technical efficacy words

  • Improve reliability
  • Improve use value

Organic light emitting diode display

ActiveUS20070008268A1improve reliability
Owner:LG DISPLAY CO LTD

Method and apparatus for correcting errors in memory device

ActiveUS20120151294A1improve reliability
Owner:SAMSUNG ELECTRONICS CO LTD

Multiple branch predictions

ActiveUS20050268075A1improve reliabilityfacilitate availability
Owner:ORACLE INT CORP

Special compound fertilizer for corn and preparation method thereof

Owner:INST OF AGRI RESOURCES & ENVIRONMENT SHANDONG ACADEMY OF AGRI SCI
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products