JPH04277873A - Features recognizing device - Google Patents

Abstract

PURPOSE:To improve the device for recognizing the features of a man. CONSTITUTION:This device is provided with an image pickup means to pickup the images of the features from the front and the side and to output a front face video signal 31 and a side face video signal 32, a binarizing means to binarize the both signals at a level between the area of the man and the background area and to output front face and side face binary signals 51 and 52, an integrating means to calculate a first profile data by integrating the picture elements of the front face binary signal laterally to the figure, a vertical range setting means to calculate a position P1 for the top of the head based on the change of the first profile data longitudinal to the figure and to calculate vertical ranges P1-P2, a horizontal range setting means to calculate a second profile data by integrating the picture of the front face binary signal in the ranges P1-P2 longitudinally and to set a horizontal range based on the change of the second profile data lateral to the figure, and a recognizing means to recognize the features based on the first and second profile data in the set ranges.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a face recognition device for recognizing a person's face.

0002

2. Description of the Related Art In security systems and the like, it is necessary to automatically identify individuals such as operators. As such techniques, personal recognition using fingerprints and personal identification using passwords are known. On the other hand, techniques for recognizing individual faces are also known, such as Sakai et al., "Analysis of Facial Images by Computer," IEICE Theory, Vol. 56-D, No. 4,
PP226-223, (1973-4) and Sakaguchi et al., "Study on person identification using isoconcentration lines of facial images", Information Processing Science, 4E-8, (1989-3)
and so on.

[0003] In addition, the eyes, mouth, etc. are extracted from a face image by image processing, their centroids are determined, and geometric techniques such as affine transformation are performed using the distance relationship between the centroids of both eyes and the centroid of the mouth. Techniques for normalizing input positions using manual means are also known. Also, as for recognition methods, there are known techniques that discriminate based on characteristics such as distance to the center of gravity and correlation.

0004

[Problems to be Solved by the Invention] However, the above-mentioned conventional technology has a drawback in that it is difficult to extract only facial information from image data. In addition, facial imaging data includes not only external shape (silhouette) information, but also the color of the eyes and nose, etc.
Since it also includes shading information, it is difficult to identify a person using a simple system.

SUMMARY OF THE INVENTION An object of the present invention is to provide a face recognition device that can easily recognize and identify a person's face using a simple system.

[0006]

[Means for Solving the Problems] A face recognition device according to the present invention is a face recognition device that recognizes a person's face, and includes an imaging unit that images a face from the front and the side and outputs a front video signal and a side video signal. a binarizing means for binarizing the front and side image signals at a level between the area of the person and the background area and outputting the front and side binarized signals; an integrating means for obtaining first profile data by integrating pixels in the horizontal direction, and a position of the top of the person's head based on changes in the first profile data in the vertical direction with respect to the person, and processing it as facial information. Vertical range setting means for setting a vertical range to be used, and image integration of the frontal binarized signal in the vertical direction within the range set by the vertical range setting means to obtain second profile data. horizontal range setting means for setting a horizontal range to be processed as facial information based on a change in second profile data in the horizontal direction; and first and second profiles for the ranges set by the vertical and horizontal range setting means. The present invention is characterized by comprising a recognition means for recognizing a person's face based on the data.

[0007]

[Operation] According to the present invention, front information and side information of a person's face are taken in as video signals, and this is binarized to obtain external shape (silhouette) information of the front and side faces of the face. This silhouette information corresponds to the unevenness of the face,
These irregularities provide important information for identifying individuals.

[0008]

Embodiments Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.

FIG. 1 is a diagram conceptually showing a method for recognizing a person's face in an embodiment of the apparatus, FIG. 2 is a diagram illustrating a method of capturing frontal and side information of a face, and FIG. 3 is a diagram of a face recognition apparatus according to an embodiment. FIG. 2 is a diagram showing functional blocks. As shown in FIG. 2, a front camera 21 is mounted on the front side of the person 1 to be identified.
, a side camera 22 is installed on the side, and a front video signal 3 is installed.
1 and a side image signal 32 are generated. The front video signal 31 and the side video signal 32 are inputted to a front binarization circuit 41 and a side binarization circuit 42, respectively, as shown in FIG. 3, where they are compared with a reference level. Here, since the reference level is set to be intermediate between the level of the person 1 portion and the background portion, the front video signal 31 and the side video signal 32 are binarized into the person portion and the background portion. That is,
A front binarized signal 51 and a side binarized signal 52 having silhouette information as shown in FIG. 3 are obtained.

In FIG. 3, a profile data generation section 61, a range setting section 62, and a recognition processing section 63 are realized as software in, for example, a microprocessor. First, the front binarized signal 51 is given to the profile data generation section 61, where pixel integration is performed in the horizontal direction in FIG. 1, and one-dimensional profile data is generated. Then, since this profile data suddenly changes at the position P1 of the top of the head of person 1 (FIG. 1), this position P1 is first determined, and then a predetermined width (vertical width of the face) is calculated from the position P1 of the top of the head. The position P2 (FIG. 1) of the lower part of the face that has moved downward by the amount shown in FIG. Based on the position P1 of the top of the head and the position P2 of the lower part of the face, the range setting section 62 sets a vertical range (P1 to P2) to be recognized. The profile data at P1 and P2 becomes the one-dimensional profile data 71 to be recognized as shown in FIG. This range data of P1 to P2 is returned to the profile data generation unit 61 in FIG. One-dimensional profile data 72 is determined (FIG. 1). Here, since the one-dimensional profile data 72 changes suddenly at the left and right ends of the face, the width (P3 to P4) and center position of the face are determined.

Regarding the recognition of the side faces of the face, since the front side is more important than the back side (the back of the head), only the front side is extracted from the center position of the side face binarized signal 52, and the profile data generation unit 61 generates a one-dimensional profile. data 73
is required. Here, the one-dimensional profile data 73
Regarding the upper and lower widths of , the previously determined range P1 to P2 is referred to (FIG. 1). As described above, silhouette information 81 based on horizontal one-dimensional profile data 71 of front binarized signal 51, silhouette information 82 based on vertical one-dimensional profile data 72 of front binarized signal 51, Silhouette information 8 based on horizontal one-dimensional profile data 73 of side surface binarized signal 52
3 is obtained. Moreover, these silhouette information 81,8
No. 2 and 83 contain unevenness information of only the face, so face recognition can be easily and accurately performed. For specific face recognition, a neural network may be used. In other words, profile data such as the one described above may be input to the neural network and the neural network may be trained.

[0012] As described above, faces can be recognized using silhouettes, but in order to improve recognition accuracy, the following configuration may be added. That is, based on the information regarding the face range (P1 to P2 in the vertical direction and P3 to P4 in the horizontal direction) obtained by the above system,
Identify the range to be recognized and process the data pixel by pixel. That is, as shown in FIG. 4, by classifying the pixels in the recognition range into four types of circles 1 to 4 (see (a) and (b) in the same figure) and grouping them together ((c) in the same figure), 4 as shown in Figure 5
Divide and input into several neural networks. The result is then determined by majority vote. In this way, the number of pixels can be reduced by dispersing only the mask areas (P1 to P2, P3 to P4) in the image for image processing.

[0013]

Effects of the Invention As described above in detail, the present invention captures front information and side information of a person's face as a video signal, and binarizes this to provide information on the external shape (silhouette) of the front and side faces. I'm looking for. This silhouette information corresponds to the unevenness of the face, and this unevenness is important information for identifying an individual. Therefore, an object of the present invention is to provide a face recognition device that can easily recognize and identify a person's face using a simple system.

[Brief explanation of the drawing]

FIG. 1 is a conceptual diagram showing a method applied to a face recognition device according to an embodiment.

FIG. 2 is a conceptual diagram showing the capture of front and side information.

FIG. 3 is a functional block diagram showing the configuration of a face recognition device according to an embodiment.

FIG. 4 is a conceptual diagram illustrating a method of a modified example.

FIG. 5 is a configuration diagram of an apparatus that processes the method of FIG. 4;

[Explanation of symbols]

21...Front camera 22...Side camera 31...Front video signal 32...Side video signal 41... Front binarization circuit 42...Side binarization circuit 51...Front binary signal 52...Side binary signal 61...Profile data generation unit 62...Range setting section 63...Recognition processing unit

Claims (3)

[Claims] 1. A face recognition device that recognizes a person's face, comprising: an imaging unit that images the face from the front and the side, and outputs a frontal image signal and a side image signal; a binarizing means for binarizing at a level between the front and background areas and outputting front and side binarized signals; and an integration means for obtaining profile data of the person, and a position of the top of the head of the person based on changes in the first profile data in the vertical direction with respect to the person, and setting a vertical range to be processed as facial information. vertical range setting means; image integrating the frontal binarized signal in the vertical direction within the range set by the vertical range setting means to obtain second profile data; horizontal range setting means for setting a horizontal range to be processed as facial information based on a change in the profile data; and based on the first and second profile data of the range set by the vertical and horizontal range setting means. , and recognition means for recognizing a person's face. 2. The recognition means includes the first and second profile data obtained from the frontal binarized signal;
2. The face recognition device according to claim 1, wherein the face recognition device is configured to recognize a person's face based on first profile data obtained from a signal corresponding to the front half of the face of the side surface binarized signal. 3. The front view or video binarized signal in the range set by the vertical and horizontal range setting means is inputted to a learning recognition means having a neural network for each pixel or for each of a plurality of adjacent pixels. 3. The face recognition device according to claim 1, further comprising means configured to recognize a face of a person.