JPH04100178A - Navel position detection device for fingerprint - Google Patents

Abstract

PURPOSE:To easily extract an essential part as a feature point and to improve matching accuracy by determining an essential position by collation with a partial image which is made to correspond to a concentric rise line pattern. CONSTITUTION:A mask 30 consists of plural partial images which are made to correspond to one or plural upper half fingerprint rise line patterns, and binary data, coordinate data, etc., are registered previously. A mask matching part 5 while scanning a specific partial image of the mask 30 on the fingerprint image performs the pattern matching to retrieve a candidate point for positioning, moves the respective partial images in parallel according to the quantity of movement to the detected candidate point, and performs pattern matching at its periphery. An essential position determination part 6 determines the essential position according to the mask position on the fingerprint image meeting specific matching requirements. Consequently, the essential position of the fingerprint can be detected by the easy method.

Description

[Detailed Description of the Invention] (Summary) The present invention relates to a fingerprint navel position detection device, and an object of the present invention is to detect the navel position using a simple method. The mask matching section includes a mask in which a plurality of partial images cut out from the fingerprint are arranged, a mask matching section that performs pattern matching between the fingerprint image and the mask, and a navel position determination section that determines the navel position of the fingerprint. The section scans a predetermined partial image of the mask over the fingerprint image and performs pattern matching to search for candidate points for positioning, and translates each partial image based on the amount of movement to the detected candidate point. The navel position determination unit determines the navel position from the mask position on the fingerprint image that satisfies a predetermined matching condition as a result of the pattern matching, The eyebrow position is detected from the fingerprint image by comparing it with the mask corresponding to the ridge pattern of the upper half.

[Industrial application field]

The present invention relates to an improvement in a fingerprint navel position detection device.

Fingerprint matching devices take advantage of the fact that the distribution of branch points and end points of fingerprint ridges (a series of convex parts) differs from person to person.
In advance, binary images near the minutiae points of these fingerprints are cut out as windows, and the relative positional relationship between the windows and the binary images in the windows are registered, and these partial images and Identity verification is performed by pattern matching with the input image.

In order to improve the verification accuracy, it is necessary to incorporate the "Uke" part, which is the central part of the fingerprint, as a feature point, and a simple and precise position detection device is required.

[Conventional technology]

FIG. 5 is a diagram showing an example of a fingerprint image, and FIG. 6 is an explanatory diagram of a conventional center point detection method.

Figure 5 shows an example of a fingerprint image, and shows that the ridge pattern has a shape close to a concentric ellipse centered on 'MJ, and is distributed with one end point and one branch point. .

From this fingerprint image, branch points and end points as feature points are extracted using a feature extraction mask, etc. However, the knee bone formed by the end points etc. cannot necessarily be extracted as a feature point unless the sea bream position is detected. is not limited.

FIG. 6 shows an example of a method for detecting the center point (knee bone) of a fingerprint in order to set the origin coordinates of the fingerprint image (for example, Information Management VOL 26 No. 41983).

First, the ridge pattern in the fingerprint image is thinned as shown in FIG. 6, and is represented by the center line (core line) of the book. A point S located above the screen among the plurality of skeleton patterns. Starting from So, move one skeleton down from So, and trace the skeleton left and right to find the vertex S1. Next, move one core line down from S and find the vertex S2. In this way, the vertices of the skeleton are sequentially determined, and the apex P when the skeleton cannot be lowered is taken as the center point.

[Problem to be solved by the invention]

The conventional fingerprint center point detection method described above is a method of tracing the core line toward the center point of the fingerprint, but when thinning the line, it completely corrects ridge patterns such as short circuits due to sweat and breaks due to dryness. The process must be kept in advance, and processing takes time with a simple registration verification device.

Furthermore, since the actual position of the navel is point X in FIG. 6, even if the window is cut out centering on point P, the chain portion will not necessarily be included.

SUMMARY OF THE INVENTION In view of the above-mentioned problems, it is an object of the present invention to provide a fingerprint position detection device that detects a fingerprint position using a simple method.

[Means to solve the problem]

In Fig. 1, a diagram showing the principle of the present invention, 30 is a mask, which is made up of a plurality of partial images corresponding to one or more fingerprint ridge patterns on the upper half, and whose binary data, coordinate data, etc. are registered in advance. Ru.

5 is a mask-shaped ((at the joint part, a predetermined partial image of the mask 30 is scanned on the fingerprint image, pattern matching is performed to search for a candidate point for positioning, and the amount of movement to the detected candidate point is calculated. Based on this, each partial image is translated in parallel and pattern matching is performed around it.

Reference numeral 6 denotes a navel position determination unit, which determines the position of the navel based on the mask position on the fingerprint image when the predetermined verification conditions are not met as a result of the verification performed by the mask verification unit 5.

C) The ridge pattern of a fingerprint has a shape close to concentric circles. For this reason, a plurality of partial images corresponding to one or more ridge patterns near the armpit are registered as the mask 30.

The mask matching unit 5 performs pattern matching while scanning a predetermined partial image among the plurality of partial images on the fingerprint image to search for a candidate point for alignment, and when a candidate point is detected, it searches for a candidate point for alignment. Each partial image is translated in parallel by calculating the amount of movement, and pattern matching between the partial image and the fingerprint image is performed around each partial image.

Then, when the matching condition is satisfied at the candidate point, the navel position determination unit 6 determines the navel position from the mask position on the fingerprint image.

Note that if there are multiple candidate points that satisfy the matching conditions, the candidate point closest to the center is adopted.

Further, the partial image may be cut out from an actual standard ridge pattern (or may be generated using a pseudo ridge pattern).

As described above (the position is determined by pattern matching using a partial image corresponding to the ridge pattern of the upper half), the navel position can be easily detected and can be used as a feature point.

〔Example〕

Embodiments of the present invention will be described in detail using FIGS. 1 to 4.

Fig. 1 is a principle diagram of the present invention, Fig. 2 is a configuration diagram of an embodiment,
FIG. 3 is a flowchart of the navel position detection process, and FIG. 4 is a diagram showing the navel position detection procedure.

FIG. 2 shows an example of a fingerprint registration and verification device to which the present invention is applied, including a numeric keypad 1 for inputting registration/verification instructions, a registration number, etc., and a fingerprint for detecting a fingerprint image from a finger 20 in contact during registration and verification. It is composed of an image input device 2 and a registration verification section 3.

In the registration verification unit 3, 4 is a registration unit that extracts a plurality of feature points from the registration screen 32 (corresponding to the fingerprint image 31) stored in the screen information storage unit 11, and cuts out a window containing the feature points. , binary data in the window, coordinate data of the window, etc. are registered in the window information storage unit 10 as window information, and the mask matching unit 5. The step 1s detected by the lung position determination unit 6 is extracted and registered as a feature point.

Reference numeral 5 denotes a mask matching section, which matches the registered mask 30 and the registration screen 32 by a matching process to be described later, and notifies the mask position determining section 6 of a mask position that satisfies a predetermined matching condition.

Reference numeral 6 denotes a center position determination unit which determines the navel position by calculating the interval between each ridge pattern, the relative position of the good intention, etc. from the notified mask position.

Reference numeral 8 denotes an image input unit that binarizes the fingerprint image data input from the fingerprint image input device 2 and stores it in the screen information storage unit 11 as screen information having coordinates.

Reference numeral 7 denotes a verification unit, which verifies the identity of the person by comparing the input fingerprint image with the registered window.

As shown in FIG. 1, the mask 30 is made by selecting each ridge pattern from two ridge patterns closest to the position S out of an actual ridge pattern or a pseudo ridge pattern taken from a finger of standard thickness. Three partial images are cut out into a window shape, and the coordinate data of the windows ① to ② and the binary data in the window are registered in the mask storage section 9 as a mask 30. At this time, window ■ is defined as a window for positioning.

Note that a guide is provided on the contact surface of the fingerprint image input device 2 that the finger 20 touches, so that the detected fingerprint image has an inclination within ±5° on the screen.

Therefore, pattern matching between the mask 30 and the registration screen 32 is performed almost accurately.

In the above configuration, navel position detection processing as shown below is performed when extracting feature points. (Figure 3.

(See FIG. 4) (1) When the registration screen 32 is stored in the screen information storage section 11 by the image input section 8, the mask matching section 5 uses the window ■ set as the alignment window as the registration screen 32. The registration screen is scanned in a spiral pattern starting from the coordinates P1 given in the window (2) above, and pattern matching is performed at each position. Figure 4 (1) (2) Then, calculate the degree of inconsistency at each position, verify whether the degree of inconsistency is less than the threshold value, and if it is less than the threshold value, the point [for example, in Figure 4 (■ ).

A movement vector A is calculated using S in (III) as a candidate point.

(3) Move Goodwill ■～■ Move in parallel by Hex I and Ru A,
Scan each window ■~■ within a predetermined range around it,
Perform pattern matching. FIG. 4 (IV) When a predetermined matching condition is passed, for example, if there are more than a predetermined number of windows below the threshold value within a predetermined range, the position [p++ vector A , P2 + hector (A + α1
)... Record j.

Similarly, the next candidate point is searched and matching processing is performed.

Perform pattern matching for each candidate point as in step 2,
The shoulder position detection unit 6 collects position information of candidate points that have passed the matching conditions.
Notify.

In addition, there are multiple candidate points that passed, and
), the lowest one is notified to the key position determination unit 6 as shown in S3.

(4) The navel position determination unit 6 is predetermined by a predetermined determination method, for example, the distance between the ridge patterns, that is, the relative positional relationship between the windows ■ and the window ■, and the positional relationship between the windows ■ and window ■. The position is determined as position S.

FIG. 4 (V) When the first place ZS is determined as described above, the registration unit 4 checks the presence or absence of end points etc. using a feature extraction mask, etc., and then selects the state place ZS.
Extract and register the window containing the .

As described above, positioning is performed using one of the partial images corresponding to the concentric ridge pattern, and pattern matching is performed by moving the other partial image and scanning its surroundings, regardless of individual differences. , one can easily determine its first place.

12i makes it easier to incorporate the focal point ys as a feature point, making it possible to improve matching accuracy.

[Effects of the Invention] As explained above, the present invention determines the bladder position by comparing it with a partial image corresponding to a concentric ridge pattern, and the station area can be easily extracted as a feature point. The effect of improving matching accuracy is significant.

[Brief explanation of drawings]

Fig. 1 is a principle diagram of the present invention, Fig. 2 is a configuration diagram of an embodiment,
FIG. 3 is a flowchart of the sugar position detection process, FIG. 4 is a diagram showing the navel position detection procedure, FIG. 5 is a diagram showing an example of a fingerprint image, and FIG. 6 is an explanatory diagram of the conventional center point detection method. . In the figure, ■ is a numeric keypad, 2 is a fingerprint image input device, 3 is a registration verification section, 4 is a registration section, 5 is a mask verification section, 6 is a side position determination section, 7 is a verification section, 8 is an image input section, 9 is the mask storage section,
10 is a window information storage unit, 11 is a screen information storage unit, 20 is a finger, 30 is a mask, and 32 is a registration screen. 31 is a fingerprint image; Principle diagram of the present invention; Figure 1; Figure 2; Flowchart of moxibustion extraction process; ) Diagram representing candidate points (rV) Diagram representing bona fide scanning (v) Diagram representing an example of determining the seed W Diagram representing the abandoned device detection procedure Diagram

Claims (1)

[Claims] A mask (30) in which a plurality of partial images cut out from one or more fingerprint ridge patterns of the upper half of the fingerprint are arranged, and between the fingerprint image (31) and the mask (30). The mask matching unit (5) includes a mask matching unit (5) that performs pattern matching, and a navel position determining unit (6) that determines the navel position of the fingerprint.
5) searches for candidate points for positioning by scanning a predetermined partial image of the mask over the fingerprint image, performs pattern matching, and parallelizes each partial image based on the amount of movement to the detected candidate point. The navel position determining unit (6) determines the navel position from the mask position on the fingerprint image that satisfies a predetermined matching condition as a result of the pattern matching. A fingerprint navel position detection device, characterized in that the navel position is detected from a fingerprint image by comparing it with the mask corresponding to the ridge pattern of the upper half of the fingerprint.