JPS60123970A - Picture reader - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION (a) Technical field to which the invention pertains The present invention relates to an image reading device, which scans and processes high-speed images of gray scale images, such as fingerprint images, and performs image retrieval. The present invention relates to an image reading device that creates data.

(b) Conventional technology Fingerprints are said to be the same for everyone and remain unchanged throughout life, and have been used for many years to identify people around the world. Until now, fingerprint verification has been done manually, but in recent years Automatic fingerprint identification systems have been put into practical use due to advances in pattern recognition technology and computer technology.

This type of system is comprised of an input system that manages fingerprint reading, a verification system that verifies the fingerprint reading results, and a search system that searches for an image of the corresponding fingerprint based on the verification results. In order to accurately identify a person, the matched fingerprints are visually compared and confirmed with the searched fingerprint image. Conventionally, the fingerprint data registered in the search system was data with shading. For this reason, the amount of data that is registered in the search system becomes enormous. For example, fingerprint data with one shade is 512X64
0x4 people = 160, (with it, J1.6 per person
M byte), ioo million people's W record is actually 1,60
0 GB capacity. One optical disk, which is currently the highest density recording medium, has a capacity of just over IG bytes, but even if optical disks are used, the number of disks will be more than 1,500. Furthermore, in order to shorten the search time, it is necessary to access a plurality of optical disks at the same time, which increases the size of the system.

When magnetic disks are used, their recording density is less than one-tenth that of optical disks, so the disadvantage is that the number of magnetic disks and the computer system that controls them are even larger.

(C) Purpose of the Invention The present invention has been made in order to eliminate the above-mentioned drawbacks of the conventional technology.Therefore, an object of the present invention is to provide a new function for data compression of visual fingerprint image data. It is in.

Another object of the present invention is to provide a novel fingerprint image reading device having a compression function.

Still another object of the present invention is to provide a novel image reading device having functions of reading fingerprints, extracting features, and compressing fingerprint data.

(d) Structure of the Invention According to the present invention, there is provided a first means for correcting input image data according to the characteristics of the input means, and a processing power of the first means is applied to the geometry of the image data. a second means for processing using the geometrical characteristics; a third means for inputting the processing output of the second means and extracting image features; and a third means for inputting the processing output of the second means and performing geometric a fourth means for encoding the output of the fourth means using the geometrical properties of the image data; and a sixth means for inputting the output of the means and outputting it to an external information processing device. Thus, an image reading device is obtained which is characterized in that it outputs significantly less image data.

(e) Embodiment of the Invention Next, a preferred embodiment of the present invention will be specifically explained with reference to the drawings.

Figure @1 is a system diagram of an automatic fingerprint identification system using the image reading device of the present invention. In fig.

The input system 10 includes a computer 11. which performs system control. An image reading device 12 that reads the fingerprint card 14 on which a fingerprint is stamped. An image reading device 12' installed at a remote location, and a remote control device 13 for connecting the image reading device 12' and the computer 11. Other input/output devices (MT,
Feature data extracted from the fingerprint is transferred to the verification system 20, and image search data obtained by compressing the fingerprint image is transferred to the search system 30.

The verification system 20 includes a computer 21. which performs system management. It consists of a magnetic disk group 22 in which fingerprint characteristic data is stored, and a verification device 23 that collates the fingerprint characteristic data input from the input system 10 and the fingerprint characteristic data stored in the magnetic disk 22. Verification, notification of verification results to other systems, and registration of fingerprint characteristic data to be added to the new nine.

The search system 30 includes a computer 31 that performs system control, an optical disk 32 that records compressed image search data, an image printer 33 that decompresses and prints the compressed image search data, and other input/output devices (Fig. (not shown), registers compressed image search data from the input system 10, and prints out the corresponding fingerprint based on the verification result from the verification system 20.

FIG. 2 shows an embodiment of the present invention. The fingerprint image stamped on the fingerprint card 100 can be printed on, for example, a lamp, a lens, a CC
It is read by a scanner 101 consisting of a D sensor and a feed mechanism (not shown) and converted into scan data 101S. Since the scan data 101S includes uneven illuminance of the illumination lamp and uneven sensitivity of the CCD sensor, the input processing unit 102 corrects these and also corrects uneven density in the stamping state.

After removing noise from the image data and emphasizing the image data, the image processing unit 103 extracts the ridges of the fingerprint by making full use of the geometric properties of the fingerprint.

This is because most of the fingerprint impression is composed of band-like patterns, so the small areas of the fingerprint impression have a certain direction. The feature extraction unit 104 detects end points and branch points of ridges, which are feature points of a fingerprint.

The reduction processing unit 105 reduces the fingerprint data from which the ridges have been extracted by the image processing unit 103. Methods of reduction include a simple thinning method and a method of interpolating data between pixels. The amount of data can be easily reduced to K by setting the length and width of the two-dimensional image to percentages, for example.

The data encoding unit 106 further reduces the amount of data by encoding the data. There are CCITT standard MH method and MR method for data encoding, but if you use the predictive gradient encoding method that predicts the appearance of data using the geometric properties of fingerprints, further compression is expected. can. Therefore, the amount of data can be significantly reduced by reduction processing and data compression processing. The interfaine section 107 is
In the feature extraction unit 104, J:! The extracted fingerprint feature data and the image search data encoded by the data encoding unit 106 are input and transferred to a higher-level control device (not shown). In this embodiment, a configuration including the data encoding section 106 has been described, but a configuration not including the data encoding section 106 is also conceivable, but in this case, the data compression rate will not be improved.

FIG. 3 shows a typical image of a fingerprint according to this embodiment. FIG. 3(a) is an image output from the input processing unit 102, and the amount of data is 512 x 640 x 4 bits = 1
60 bytes. FIG. 3(b) shows the image processing unit 103.
The amount of data in the output image is 512 x 640
X 1 focus = 40 bytes, Figure 3 (C) is an image output from the reduction processing unit 105, and the amount of data is 256
X 320 X 1 pinto = 10 bytes. Conventionally, Fig. 3(a) has been used as image search data, but by using Fig. 3(c), the data can be reduced to JIIOK bytes/160 bytes without encoding processing. I can do it.

Of course, the image quality of No. 31N(C) is not worse than the image quality of the original image of FIG. 3(a).

(f) As described in detail, according to the present invention, data after image processing based on the geometrical properties of a fingerprint is reduced. ,
By performing imi-J-encoding processing, it is possible to significantly reduce the amount of data while maintaining the quality and feather characteristics of the original image, which is highly effective in downsizing large-capacity image retrieval systems.

[Brief explanation of drawings]

Fig. 1 is a system diagram of an automatic fingerprint identification system, Fig. 2 is a block diagram showing an embodiment of the present invention, and Fig. 3 (3), (
b) and (C) are diagrams showing one typical fingerprint stroke in this embodiment. 10...Person cuff stem, 11...Computer, 1
2,121. ...Image reading device, 13...Remote control device, 14...Fingerprint card, 20...Verification system, 21...Computer, 22...Magnetic disk, 23...Verification device, 30. ...Search system, 31
...Computer, 32...Optical disc, 33...
・Image printer, 100... Fingerprint card, 101
... Scanner, 102 ... Input processing section, 103.
...Image processing unit, 104...Fat feature extraction unit, 105...
・Reduction processing unit, 106...Data encoding unit Patent applicant: NEC Corporation Representative: Yuta Kumagai, patent attorney Figure 1 Figure 2 (0) (b) Figure 3 (C)

Claims (1)

[Claims] an input processing means that corrects the input image data according to the bamboo nature of the input means; an image processing means that inputs the processing result of the input processing means and processes it using the geometric properties of the image data; feature extraction means for extracting features of an image based on the processing results of the image processing means; reduction processing means for geometrically reducing the processing results of the image processing means; and reduction processing means for geometrically reducing the processing results of the reduction processing means. data encoding means for encoding data using properties; and output means for inputting the processing results of the feature extraction means and the processing results of the data encoding means and outputting them to an external information processing device. An image reading device that is equipped with an image reading device.