Image analysis – Applications – Personnel identification
Reexamination Certificate
2001-08-01
2002-06-11
Mehta, Bhavesh (Department: 2621)
Image analysis
Applications
Personnel identification
Reexamination Certificate
active
06404903
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a system for identifying individuals by comparing an input image with a dictionary image previously stored.
2. Related Art
A technology is well known by which an individual is identified from physical information (about the face, iris and fingerprints, for example) of human beings and other animals. To cite an example, this technology is disclosed in U.S. Pat. No. 5,291,560.
In the technology revealed in the above literature, the image of the iris of an eye of a human being is converted into codes of 0 and 1, and the codes are stored as a dictionary. In an actual identification process, an input iris image is converted into codes of 0 and 1, and the codes are compared with codes stored in the dictionary to establish the identity of an individual.
By the conversion of an input image into codes of 0 and 1, the amount of processing can be reduced, which leads to savings of storage capacity. In the identification of an individual by fingerprints, too, it is ordinary to detect the features from an input image, and store only data on the features as a dictionary. Likewise, in the identification of an individual by a face image, the amount of information is reduced by the use of mosaic images or the facial features.
In the conventional technology mentioned above, however, the input image is converted into codes of 0 and 1 or information about the facial features. Therefore, if the system should make an error in understanding data, it is difficult for a user without specialized technical knowledge to find the cause of the problem, the system will be a burden for the user to deal with.
To take an identification system using the face as an example, if a smiling face is registered when a dictionary is created and a serious face is input in the actual operation of the system, the system may mistake this as a discrepancy in the comparison step. On the other hand, if days passed from the registration till the identification operation and the user forgot his or her facial expression on the day of registration, it is impossible to conduct identification by simply displaying the registered image and the input image for comparison, with the result that it is impossible to decide whether the identification itself is correct or wrong.
In a system which has images at the time of dictionary creation stored in advance, if a misidentification occurred, even when the stored image at issue is displayed, it takes time until you can find that region from the image of the whole face which was conducive to the misidentification.
Accordingly, it has long been desired that a system should be developed which enables the user to find the cause of the misjudgment even when misidentification occurred.
With a system for identifying individuals by the prior art mentioned above, when this system is applied to identification of animals, such as dogs, cats or horses, a problem arises as follows.
Animals are different from human beings and you cannot expect them to behave in a manner favorable to the system. For example, when taking a picture of the iris of the eye with a video camera, a person is at a standstill squarely facing the camera. However, in the case of an animal, it does not look at the camera, and the face is moving most of the time.
Therefore, in the identification of animals, the images taken are inferior in picture quality, which is often responsible for a failure of identification. For this reason, it was necessary to take images repeatedly.
In this respect, it has been expected that a system for identifying individuals should be created which can identify an individual even when the image obtained does not have a good picture quality.
To solve this problem, the present invention adopts the following arrangement.
SUMMARY OF THE INVENTION
According to a first aspect of the invention, a system for identifying individuals comprises:
an image memory means for storing an input image of a individual as an object to be identified;
a dictionary for having data on the features of collation objects stored in advance;
a dictionary image memory means for storing dictionary images as a basis on which to extract data on the features of the collation objects;
an individuals identification means for analyzing the input image held in the image memory means and comparing said data on the features stored in the dictionary to thereby identify the individual;
an identification result analysis means for analyzing the identification result by the individuals identifying means and detecting that area of the region used for identification in the input image which does not agree with the dictionary data;
a result output control means for, according to identification result by the individuals identifying means, issuing an analyze command to the identification result analysis means and, according to analysis result by the identification result analysis means, issuing a command to display the discordant portion to the identification result analysis means, and deciding whether or not to display the input image and the dictionary image;
an image output control means for, according to a result of decision by the result output control means, controlling display of the discordant portion, the input image and the dictionary image; and
a display for displaying an image outputted by the image output control means.
With regard to the first aspect, the input image and the dictionary image are images of the iris of the human eye, and the individuals identifying means conducts identification by comparing iris codes. However, any arrangement other than was described above may be applied so long as it identifies an individual.
With continuing reference to the first aspect, the identification result analysis means detects an area where the input image does not coincide with the dictionary image by comparing iris codes, for example. The result output control means makes a decision to have at least where the discrepancy occurred displayed according to an analysis result of the identification result analysis means.
By this operation of the first aspect, even if the identity of the person is not established, where the discrepancy occurred is displayed, so that it is possible for the user to easily guess the cause of misjudgment.
It may be provided that, if the input image agrees with the dictionary data to such an extent that the identification result of the individuals identifying means is larger than a predetermined threshold value, the result output control means makes a decision not to issue an analyze command to the identification result analysis means nor issue a display command to the image output control means.
The rationale is that, if the input image agrees with the dictionary data to such an extent that the identification result of the individuals identifying means is larger than the predetermined threshold value, this means that the person is identified as a correct person. In this case, the system does not display analysis of the identification result nor display the input image. Therefore, unnecessary processing is not executed and a reduction of processing load can be expected.
It may be provided that, if the input image agrees with the dictionary data to such an extent that the identification result of the individuals identifying means is higher than the predetermined first threshold value but the amount of agreement is smaller than a second threshold value, which is smaller than the first threshold value, the result output control means makes a decision not to issue an analyze command to the identification result analysis means nor issue a display command to the image output control means.
Here, the analyze process and the image display process are not executed when the person is identified as a correct person or the person is identified as somebody else or the quality of the input image is judged to be inferior, and the analyze process and the image display process are executed only when the identity of the person is uncertain. The cases where
Kuno Yuji
Okano Kenji
Mehta Bhavesh
Oki Electric Industry Co, Ltd.
Venable
Voorhees Catherine M.
LandOfFree
System for identifying individuals does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System for identifying individuals, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System for identifying individuals will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2902794