TWI460666B - Anti-forgery fingerprint verification system - Google Patents

Description

Anti-counterfeiting fingerprint identification system

The invention combines the traditional fingerprint identification with the vein identification, so that the user can also take out the vein image by the vein sensor while the finger is captured by the finger image. Since the invention can further match the traditional fingerprint identification result with the vein identification, the effect of increasing the recognition rate and preventing the fake fingerprint or the finger being cut can be successfully deceived by the traditional fingerprint identification system.

Due to the increasing progress of electronic technology, it has become increasingly unsafe to use personal passwords as the basis for identity identification. Therefore, more and more rigorous identification methods, such as face recognition, pupil recognition, retina recognition, fingerprint recognition and vein identification, have been developed in recent years. Because the user's fingerprint is the most convenient, in recent years, products such as mobile phones, notebook computers, personal computers, and ATM cards have gradually replaced fingerprints with traditional passwords. However, in the higher-level security agencies, there is often only one layer of identification methods that are not very secure. Therefore, we use an anti-counterfeiting fingerprint identification method combining fingerprint identification mechanism and vein identification mechanism to improve the security and reliability of the system.

The identification system is based on fingerprint recognition, with an auxiliary vein identification mechanism to ensure that the user who inputs the fingerprint is actually obtained from the certifier's finger, not from a duplicated fingerprint or one. Taken on the cut finger.

The invention utilizes the characteristics that the user needs to place the finger on the fingerprint sensor when performing fingerprint recognition, and the fingerprint sensor used in the traditional fingerprint identification is matched with the vein sensor, so that the fingerprint sensor can capture the fingerprint image while allowing the fingerprint sensor to capture the fingerprint image. The vein sensor takes out the vein image of the user. After obtaining fingerprints and vein images, the central processing unit in the system extracts the characteristics of the fingerprint image captured by the sensor and the characteristics of the vein image, and the authenticator stored in the storage (database). Compared with the characteristics of the veins and the vein characteristics, if the extracted fingerprint features meet the fingerprint characteristics of the authenticator and the comparison score of the vein features exceeds the preset threshold, the system authenticates by allowing the controller to operate; If the system authentication fails, the controller is not allowed to operate.

The preferred method of the present invention is described in detail below in conjunction with the drawings.

As shown in Figure 6, the fingerprint recognition is to put the hand on the hardware, wherein the fingerprint sensor 61 and the NIR CCD 62 can acquire the fingerprint and the original image of the vein, and the others use the 8051 IC and RS232 circuit to transmit the image to the computer. And then processed by the computer's central processing unit.

Figure 1 is the main method of implementation of the present invention. The left side of Figure 1 is the process of establishing the user database storage, and the right side of Figure 1 is the identification process. As shown in Figure 1, the processes 3 to 9 and the processes 16 to 22 are almost In the same way, after processing the image in the same process, the FSVM classification method is used to identify it. Therefore, we only describe the process of 3~9. First, enter the user's name 2 first, then use the NIR CCD Capture the image of the vein 3, use the fingerprint sensor 4 (MBF200) to obtain the fingerprint image, and then find the registration point 5 of the fingerprint.

The fifth picture is an example of a fingerprint registration point found. Then, the enhanced finger vein image 6 is taken out, the fourth figure is the process of enhancing the finger vein image, 45 is the original vein image input 41, and then subtracted by the average box filter 47 and the low-pass filter 48, and then reused. A suitable threshold value 49 separates the vein from the background 43, then eliminates the edge portion 50, and eliminates the noise enhanced vein image 44 by normalizing 52 (Erosion) and expanding 53 (Dilation).

The progressive fingerprint enhancement process 7 is as shown in the second figure. First, the original fingerprint image is broken 27, and the contrast of the fingerprint image is enhanced by using the histogram equalization 28, and then the Sobel operator is used to calculate the direction of the image. The direction of the image is smoothed by 29, and the fingerprint and the background are segmented 30, and then the noise 33 is filtered by the Gabor filter to smooth the direction 34 of the image so that the fingerprint no longer has a broken block. The third picture is an example of image after fingerprint enhancement, 37 represents the original fingerprint image, 38 represents the fingerprint and the background, and 39 represents the enhanced image, and there is no broken piece like 37.

Finally, the fingerprint feature extracted and the fingerprint feature of the database stored by the database are compared 23 to determine whether it is the same user. If yes, proceed to the next program to determine whether the feature comparison score of the captured vein image is higher than the threshold value of 24. If not, the user fingerprint is required to be retrieved again. If the vein characteristics are compared with the venom characteristics of the authenticator of the database, the comparison score is higher than the threshold and the certification is passed 25, otherwise the certification is not passed26. If it is applied to a general access control system, Whether it is certified or not can be regarded as whether the access control system can be opened; if it is applied to a flash drive or a hard disk, it can be regarded as whether the data on the portable disk or the portable hard disk can be read.

1‧‧‧Users establish database process

2‧‧‧Users register new ID English name

3‧‧‧Using infrared rays to obtain vein original images

4‧‧‧Get fingerprint original image

5‧‧‧Identification of fingerprint registration point and estimation of vein registration point offset

6‧‧‧Get processed finger vein images

7‧‧‧Enhanced enhanced fingerprint imagery

8‧‧‧Using wavelet transform, Principal component analysis (PCA) is used to reduce the dimension to obtain the feature vector.

9‧‧‧Get wavelet feature vector

10‧‧‧Combined with the new ID number

11‧‧‧Combined with the new ID number

12‧‧‧written into the finger vein database

13‧‧‧written into the fingerprint database

14‧‧‧User Identification Process

15‧‧‧ Login personal ID name

16‧‧‧Get fingerprint original image

17‧‧‧Using infrared rays to obtain vein original images

18‧‧‧The location of the fingerprint registration point and the estimation of the vein registration point offset

19‧‧‧Enhanced enhanced fingerprint imagery

20‧‧‧Get processed finger vein images

21‧‧‧Get wavelet feature vector

22‧‧‧ After wavelet transform, Principal component analysis (PCA) is used to reduce the dimension to obtain the feature vector.

23‧‧‧Compatible with fingerprint database comparison

24‧‧‧Compared with the finger vein database, the correct rate is high enough

25‧‧‧Accepting user login

26‧‧‧Reject user login

27‧‧‧Enter fingerprint broken image

28‧‧‧Histogram equalization

29‧‧‧ Calculate image direction and smooth direction

30‧‧‧Divided fingerprints and background

31‧‧‧Gabor filter

32‧‧‧Vertical to smooth image orientation

33‧‧‧Gabor filter

34‧‧‧Smoothing image orientation

35‧‧‧Use the threshold value to binarize the image

36‧‧‧Enhanced output fingerprint image

37‧‧‧ Original fingerprint imagery

38‧‧‧Segmentation of fingerprints and background images

39‧‧‧Enhanced fingerprint image

40‧‧‧ Fingerprint image output after enhancement

41‧‧‧ Originally imported vein images

42‧‧‧Vascular images filtered by filters

43‧‧‧ vein images found using threshold values

44‧‧‧Intravenous images after erosion and regularization

45‧‧‧Enter original vein image

46‧‧‧ give a suitable threshold

47‧‧‧Average box filter

48‧‧‧ low pass filter

49‧‧‧ Use the threshold to find the vein

50‧‧‧Remove the edge

51‧‧‧ formalization

52‧‧‧Erosion calculation

53‧‧‧ expansion operation

54‧‧‧Fat vein image without noise

55‧‧‧Finger registration point sample image

56‧‧‧Finger registration point sample image

57‧‧‧ Fingerprint registration point sample image

58‧‧‧RS232 connector and computer transmission

59‧‧‧Hiros power supply hole

60‧‧‧8051 IC components

61‧‧‧MBF200 fingerprint sensor

62‧‧‧NIR CCD (BE-IR30)

The first figure is a flow chart of the main method of fingerprint and vein identification of the present invention.

The second figure is a fingerprint enhancement flow chart in accordance with the present invention.

The third figure is an experimental result diagram after fingerprint enhancement according to the present invention.

The fourth panel is a flow chart of vein enhancement in accordance with the present invention.

The fifth figure is a graph of the experimental results of the fingerprint registration point in accordance with the present invention.

The sixth figure is a schematic diagram of fingerprint and vein identification in accordance with the present invention.

1‧‧‧Users establish database process

2‧‧‧Users register new ID English name

3‧‧‧Using infrared rays to obtain vein original images

4‧‧‧Get fingerprint original image

5‧‧‧Identification of fingerprint registration point and estimation of vein registration point offset

6‧‧‧Get processed finger vein images

7‧‧‧Enhanced enhanced fingerprint imagery

8‧‧‧Using wavelet transform, Principal component analysis (PCA) is used to reduce the dimension to obtain the feature vector.

9‧‧‧Get wavelet feature vector

10‧‧‧Combined with the new ID number

11‧‧‧Combined with the new ID number

12‧‧‧written into the finger vein database

13‧‧‧written into the fingerprint database

14‧‧‧User Identification Process

15‧‧‧ Login personal ID name

16‧‧‧Get fingerprint original image

17‧‧‧Using infrared rays to obtain vein original images

18‧‧‧The location of the fingerprint registration point and the estimation of the vein registration point offset

19‧‧‧Enhanced enhanced fingerprint imagery

20‧‧‧Get processed finger vein images

21‧‧‧Get wavelet feature vector

22‧‧‧ After wavelet transform, Principal component analysis (PCA) is used to reduce the dimension to obtain the feature vector.

23‧‧‧Compatible with fingerprint database comparison

24‧‧‧Compared with the finger vein database, the correct rate is high enough

25‧‧‧Accepting user login

26‧‧‧Reject user login

Claims (2)

A fingerprint and vein identification device with anti-counterfeiting function, comprising: (a) a fingerprint sensor: used in the first-stage detection mechanism, when the user places his finger in the sensing area, the device can be used to capture the fingerprint image. For the user to legally use for authentication; (b) Venous sensor: After the first-stage detection mechanism is confirmed, the second phase of the vein data comparison is started, and the device is used to capture the vein image for the user to legally cross (c) Central processor: used to judge whether the captured fingerprint image, vein image and legal Authenticator data in the storage are the same; (d) Storage: used to store the fingerprint of the legal Authenticator Venous data, as well as images captured by the user when logging into the system. Fingerprint and vein image enhancement and identification method with anti-counterfeiting function, including: (a) Strengthening finger vein image: subtracting the original vein image using Average Box Filter technology and low-pass filter, and then using a suitable one. The threshold value separates the vein from the background, eliminates the edge portion, and then uses normalization and Morphology Operation to eliminate the noise to enhance the vein image; (b) strengthen the fingerprint image of the finger: the original fingerprint image, using the histogram to equalize the fingerprint The contrast of the image is enhanced. Then the Sobel operator is used to calculate the direction of the image and smooth the direction of the image. Then the fingerprint and the background are segmented. Then the Gabor filter is used to filter out the noise and smooth the direction of the image so that the fingerprint no longer exists. Broken block; (c) Identify the user's fingerprint and vein data using the FSVM classification method.