KR101718244B1

KR101718244B1 - Apparatus and method of processing wide angle image for recognizing face

Info

Publication number: KR101718244B1
Application number: KR1020150178670A
Authority: KR
Inventors: 이우균
Original assignee: 이우균
Priority date: 2015-12-14
Filing date: 2015-12-14
Publication date: 2017-03-20

Abstract

The present invention relates to an apparatus and a method for processing a wide angle image for face recognition, and more particularly, to an apparatus and method for processing a wide angle image photographed through a wide angle camera module, A processor; And a face recognition unit for comparing one or more face images extracted by the image processing unit with registered images stored in a database and outputting recognition results.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention relates to an image processing apparatus and a method for processing a wide-

The present invention relates to a method for processing a wide-angle image in a system for recognizing a face from a photographed image using a wide-angle camera.

In recent years, in the case of a key or an access card, there is a lot of room for theft, loss, duplication, etc. Therefore, the security is relatively low, and it is necessary to carry it and to memorize the password. There are many researches on personal identification technologies with high performance and convenience.

Biometric identification systems are gradually becoming commercially available as personal identification technology. Biometric identification technology is a technology that recognizes body characteristics such as fingerprints, faces, irises, and veins that are different from person to person. Such body characteristics can not be used or copied by others like a key or a password, and can be used for security fields without any risk of being deformed or lost unless the body is damaged.

Among them, facial recognition technology is not complicated as compared with biometric techniques such as fingerprint or iris recognition. In addition, it is natural that the face recognition technology is natural and does not give a sense of rejection to the user when the image of the human being is stored using a camera. Furthermore, face recognition technology can be applied not only to the security field but also to various applications in accordance with the smartphone era.

However, the conventional face recognition apparatus has a problem that the face of a child with a small height can not be recognized due to limitation of the viewing angle of the camera.
Korean Patent No. 10-1427653 (a control method of a digital image processing apparatus for performing face recognition and a digital image processing apparatus employing this method) is known as a prior art document.

An object of the present invention is to provide a face recognition system using a wide-angle camera that can effectively recognize faces of various users and a face recognition method using the same.

The wide-angle image processing apparatus for face recognition according to an embodiment of the present invention includes an image processing unit for processing a wide-angle image photographed through the wide-angle camera module and converting the wide-angle image into a planar viewpoint image, ; And a face recognition unit for comparing one or more face images extracted by the image processing unit with registered images stored in a database and outputting recognition results.

According to another aspect of the present invention, there is provided a method for processing a wide-angle image for face recognition, the method comprising: processing an image photographed through a wide-angle camera module and converting the processed image into a planar viewpoint image; Extracting one or more face images from the transformed planar view image; And recognizing a face by comparing the extracted one or more face images with registered images stored in a database.

The wide-angle image processing method may be embodied as a computer-readable recording medium on which a program for execution in a computer is recorded.

According to an embodiment of the present invention, a face is recognized from an image photographed using a wide-angle camera, so that not only an adult with a large height but also a face such as a child with a small height can be recognized at the same time.

According to another embodiment of the present invention, a plurality of faces can be detected and recognized simultaneously from a captured image using a wide angle camera.

According to another embodiment of the present invention, a wide angle image photographed through a wide angle camera is converted into a single planar view image, and a face image in the converted planar view image is extracted, The face can be correctly recognized.

1 is a block diagram showing an embodiment of a configuration of a face recognition system using a wide angle camera according to the present invention.
FIG. 2 is a view for explaining an operation of the wide-angle camera module shown in FIG. 1. FIG.
3 is a flowchart illustrating a method of processing a wide-angle image for face recognition according to an embodiment of the present invention.
4 is a perspective view showing an embodiment of the configuration of the wide angle camera module.
5 to 8 are views for explaining embodiments of a method of processing a wide-angle image photographed through a wide-angle camera.
FIG. 9 is a block diagram showing an embodiment of the configuration of the face recognition unit shown in FIG. 1. FIG.
10 is a block diagram illustrating a configuration of a face recognition system using a wide angle camera according to another embodiment of the present invention.
FIG. 11 is a block diagram illustrating an exemplary configuration of a security system using face recognition according to an embodiment of the present invention. Referring to FIG.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings, which will be readily apparent to those skilled in the art. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. In order to clearly illustrate the present invention, parts not related to the description are omitted, and similar parts are denoted by like reference characters throughout the specification.

Throughout the specification, when a part is referred to as being "connected" to another part, it includes not only "directly connected" but also "electrically connected" with another part in between . Also, when an element is referred to as "comprising ", it means that it can include other elements as well, without departing from the other elements unless specifically stated otherwise.

In this specification, the term " part " includes a unit realized by hardware, a unit realized by software, and a unit realized by using both. Further, one unit may be implemented using two or more hardware, or two or more units may be implemented by one hardware.

In this specification, some of the operations or functions described as being performed by the terminal or the device may be performed in the server connected to the terminal or the device instead. Similarly, some of the operations or functions described as being performed by the server may also be performed on a terminal or device connected to the server.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings.

FIG. 1 is a block diagram of an embodiment of a face recognition system using a wide angle camera according to the present invention. The face recognition system 10 shown in FIG. 1 includes a wide angle camera module 100, an image processing unit 110, And a face recognizing unit 120. [0033] FIG.

1, the wide-angle camera module 100 includes a wide-angle lens and an image sensor, and captures a forward image to acquire a wide-angle image including a forward object (for example, at least one person who is a face recognition object) .

The wide-angle lens is a lens having a wide-angle field of view, and can photograph a large object with a short focal length and deep focal depth, or can photograph a wide range of areas.

Referring to FIG. 2, when a general camera module 1 having a small viewing angle? 1 (for example, about 65 degrees) is used to recognize a person's face located at the front as shown in (a) The center line L1 of the lens is directed upward and only an area including an adult face of a predetermined key or more can be photographed and a child with a small height is not included in the photographed image.

For the face recognition, the face recognition system 10 should include a display (not shown) for checking whether the face to be recognized is included in the image photographed through the general camera module 1, It is possible to recognize only one face located close to the face 1 of the face.

On the other hand, when photographing forward using the wide angle camera module 100 having a large viewing angle? 2 (for example, 90 degrees to 180 degrees) as shown in (b), the centerline L1 of the lens The image can be included in the photographed image as well as an adult having a certain key height as well as a child face having a small height.

In addition, since a wide range area in front of the wide angle camera module 100 is photographed, there is no blind spot and a display (not shown) for confirming whether a face to be recognized is included in the photographed image is displayed on the face recognition system 10 And it may be possible to simultaneously authenticate faces of a plurality of people as well as a single face regardless of the distance or the key from the wide angle camera module 100. [

For this purpose, it is preferable that an image photographed through the wide-angle camera module 100 has a high image quality of Full HD or higher.

Angle camera module 100 having the wide-angle lens and the image sensor described above can photograph a wide range of areas with a large viewing angle, but an image photographed through the wide-angle camera module 100 can not be displayed due to the characteristics of the wide- Distortion may occur.

For example, an image photographed through the wide-angle camera module 100 has a radial distortion in the circumferential direction due to an extremely short focal distance, and the wide-angle lens and the image sensor are not geometrically parallel Tangential distortion may be additionally generated.

The image processing unit 110 processes the photographed image through the wide-angle camera module 100 to extract one or more facial images.

In order to correct the distortion in the wide angle image as described above, the image processing unit 110 processes the wide angle image photographed through the wide angle camera module 100 and converts the wide angle image into a planar viewpoint image 111 For example, the wide angle correction unit 111 may correct the wide angle image by converting spherical coordinates into plane coordinates of the wide angle image photographed through the wide angle camera module 100.

A specific embodiment of a method of processing the wide angle image photographed through the wide angle camera module 100 by the wide angle correcting unit 111 and converting the wide angle image into the planar viewpoint image will be described in more detail with reference to FIGS. .

The image processing unit 110 extracts a face image including one or more faces from the converted plane view image and extracts a face image by detecting one or more face regions from the image, An area detection technique can be used.

The face recognition unit 120 compares at least one face image extracted by the image processing unit 110 with a registered image stored in a database (not shown) and outputs the recognition result.

The face recognition unit 120 extracts, from the face image input from the image processing unit 110, main features of the face (e.g., distance between two eyes, length and width of the nose, length of the jaw line, The feature values of the extracted face are compared with one or more facial feature values stored in a database (not shown), and the facial features are compared with the facial feature values of the facial recognition system It is possible to authenticate whether or not it is the face of a user registered in advance.

Hereinafter, the configuration of the face recognition system 10 according to the present invention and the face recognition method according to the operation will be described in detail with reference to FIG. 3 to FIG.

FIG. 3 is a flowchart illustrating a wide-angle image processing method for face recognition according to an exemplary embodiment of the present invention. Referring to FIG. 3, Will be described in connection with a block diagram showing an embodiment. In the meantime, among the wide-angle image processing methods shown in FIGs. 1 and 2, description about the same things as those described with reference to FIGS. 1 and 2 will be omitted.

3, the image processing unit 110 of the face recognition system 10 processes a wide-angle image photographed through the wide-angle camera module 100 and converts the wide-angle image into a planar viewpoint image (step S310) At least one face image is extracted from the image (step S320).

As shown in FIG. 4, the wide-angle camera module 100 may be configured such that a wide-angle lens 101 is disposed on a camera module board 102 on which an image sensor is formed.

On the other hand, the wide-angle camera module 100 has a viewing angle of 90 degrees to 180 degrees and can capture a wide range of the front area to obtain the wide-angle image 500 as shown in FIG.

A plurality of people may be included in the wide-angle image 500 photographed through the wide-angle camera module 100, and the faces of the plurality of people may be located at different heights or at different distances.

In this case, the image processing unit 110 performs the wide angle correction on the panorama ROI (Region Of Interest) in the horizontal direction within the wide angle image 500 captured through the wide angle camera module 100, At least a portion of the corrected image 600 is converted into a plane view.

In addition, the image processing unit 110 may extract a plurality of face images 610 and 611 located at different heights or distances within the corrected image 600.

As described above, the wide angle correction unit 111 included in the image processing unit 110 corrects the distortion of the wide angle image by converting the coordinates of the wide angle image photographed through the wide angle camera module 100 into spherical coordinates to plane coordinates .

More specifically, the wide-angle correction unit 111 defines a camera model for the wide-angle camera module 100, estimates the parameters of the camera model from the calibration pattern, and then transforms the spherical coordinates of the wide- It can be converted into coordinates.

The wide-angle correction unit 111 applies various techniques in addition to the above-described method to thereby correct the distortion of the wide-angle image, which is one embodiment of the present invention. Can be corrected.

According to an embodiment of the present invention, the position of a viewpoint used by the wide-angle correction unit 111 to correct the distortion of the wide-angle image can be freely set to one or two or more, The wide angle can be corrected for each of a plurality of ROIs within the image photographed through the module 100. [

7, the wide angle correcting unit 111 corrects the wide angle of view (ROI) in the horizontal direction based on the point of time when the image 700 is positioned at the center of the image in the image 700 captured through the wide angle camera module 100, Correction can be performed to convert into a single planar view image 710. [

8, the wide-angle correction unit 111 performs a wide-angle correction on a plurality of ROIs based on a plurality of viewpoints within the image 700 photographed through the wide-angle camera module 100 It is also possible to construct a plurality of plane view image images 751 to 754 by performing wide angle correction at the same time.

The image processing unit 110 may extract one or more face images from one or a plurality of planar viewpoint images subjected to the wide angle correction as described above.

Thereafter, the face recognizing unit 120 recognizes a face by comparing one or more face images extracted by the image processing unit 110 with a registered image stored in a database (not shown) (step S330).

Referring to FIG. 9, the face recognition unit 120 according to an embodiment of the present invention includes an area detection unit 121 for detecting a plurality of feature areas from each face image, And a matching determination unit 123 that compares a plurality of points in the normalized face image with a plurality of points of a registered image stored in a database (not shown) can do.

For example, the region detecting unit 121 may detect a face region or an eye region from a face image, and more specifically may detect an eyebrow, a nose, a mouth, a jaw line, and the like in order to detect a face region or an eye region have.

On the other hand, the normalization unit 122 may normalize (or normalize) the face image based on the face region or the eye region detected by the region detection unit 121 according to predetermined conditions, Size, lighting condition, facial expression, and the like.

The matching determining unit 123 may set a key point for each of a plurality of predetermined positions of the normalized face image and may determine a key point by using a Scale Invariant Feature Transform (SIFT) or a Speed Up Robust Features (SURF) Keypoints of the face image can be set.

The matching determining unit 123 may then generate a key point descriptor for each of a plurality of key points (e.g., eyebrows, eyes, nose, and mouth) set on the face image. Specifically, To generate a key point descriptor based on a Haar wavelet response value for a region of a predetermined range that includes the < RTI ID = 0.0 >

In addition, the matching determination unit 123 may determine whether or not matching with a previously registered face image stored in a database (not shown) using a plurality of key point descriptors generated for the face image as described above.

For example, the matching determination unit 123 compares each of the key-point descriptors of the face image extracted by the image processing unit 110 with a plurality of key-point descriptors for the previously registered face image stored in the database (not shown) It is possible to judge whether or not the matching between the face images is performed.

As described above, when the image processing unit 110 extracts a plurality of facial images from the image photographed through the wide-angle camera module 100, the face recognizing unit 120 extracts a plurality of facial images It is possible to detect a face image corresponding to a pre-registered user among a plurality of face images through the comparison with a plurality of pre-registered images stored in a database (not shown).

FIG. 10 is a block diagram illustrating a configuration of a face recognition system using a wide-angle camera according to another embodiment of the present invention. FIG. 11 is a block diagram of a configuration of a security system using face recognition according to the present invention, It is a road map. The description of the same components as those described with reference to Figs. 1 to 9 among the configurations shown in Figs. 10 and 11 will be omitted below.

10, the communication module 150 provided in the face recognition system 10 includes a wide angle image photographed through the wide angle camera module 100, a plane view image processed by the image processing unit 110, At least one of the face image extracted from the face recognition unit 120 and the recognition result output from the face recognition unit 120 may be transmitted to the server 20 under the control of the control unit 140. [

The data or information transmitted from the face recognition system 10 to the server 20 may be transmitted to the user's portable terminal 30 and may be provided to the user through a screen of the portable terminal 30 or the like.

For example, when the user views the image transmitted from the face recognition system 10 installed on the front door of his / her house through the portable terminal 30 and confirms the visitor, a request to open the front door is transmitted to the portable terminal 30 ).

In this case, the communication module 150 provided in the face recognition system 10 receives the front door opening request input by the user through the portable terminal 30 through the server 20, and performs a corresponding operation can do.

Also, the communication module 150 may transmit at least one planar view image selected from a plurality of planar view images as described with reference to FIG. 8 to the outside of the server 20 or the like.

For example, when the image processing unit 110 processes a wide-angle image photographed through the wide-angle camera module 100 to construct a plurality of planar viewpoint images corresponding to each of a plurality of ROIs, By controlling the plane view image of a certain region of interest selected from the user rather than the entire image photographed through the camera module 100 or selected as a significant region by a specific algorithm to be transmitted to the server 20 through the communication module 150 , So that the data can be efficiently transmitted to the portable terminal 30 of the user.

According to another embodiment of the present invention, the control unit 140 may store the plan view image processed by the image processing unit 120 or the extracted face images in the memory unit 160 according to the time sequence.

In this case, under the control of the control unit 140, the image processing unit 110 can acquire facial feature values (e.g., information indicating impression details) for each person in the image, And stores the extracted key-level image in the memory unit 160 together with the time information.

In addition, as described above, only the plane view image of a certain region of interest selected from the user, not selected from the entire image photographed through the wide angle camera module 100, or selected as an important region by a specific algorithm, is stored in the memory unit 160 The storage capacity of the memory unit 160 can be efficiently used.

According to another embodiment of the present invention, the image tracking unit 180 included in the face recognition system 10 may perform the movement of the face image extracted by the image processing unit 110, It is possible to carry out the function of tracking within.

Meanwhile, the movement information tracked by the image tracking unit 180 may be stored in the memory unit 160 together with the time information through the controller 140.

At least a part of the face recognition method according to the present invention may be stored in a computer-readable recording medium. The computer-readable recording medium may be a ROM, a RAM, a CD -ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like, and may also be implemented in the form of a carrier wave (for example, transmission over the Internet).

The computer readable recording medium may be distributed over a networked computer system so that computer readable code can be stored and executed in a distributed manner. And, functional programs, codes and code segments for implementing the above method can be easily inferred by programmers of the technical field to which the present invention belongs.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, but, on the contrary, It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the present invention.

Claims

An image processing unit for processing a wide angle image photographed through the wide angle camera module and converting the wide angle image into a planar viewpoint image and extracting one or more face images from the planar viewpoint image;
A face recognition unit for comparing one or more face images extracted by the image processing unit with registered images stored in a database and outputting recognition results; And
And an image tracking unit for tracking movement of one or more face images extracted by the image processing unit,
And movement information tracked by the image tracking unit is stored together with time information.

The apparatus of claim 1, wherein the image processing unit
Wherein the conversion unit converts spherical coordinates to plane coordinates for the wide-angle image.

The apparatus of claim 1, wherein the image processing unit
And processing the wide-angle image to generate a plurality of planar viewpoint images corresponding to each of the plurality of ROIs.

The method of claim 3,
And a communication module for transmitting at least one planar viewpoint image selected from the plurality of planar viewpoint images to the outside.

The method of claim 3,
And a memory unit for storing at least one planar viewpoint image selected from the plurality of planar viewpoint images.

The apparatus of claim 1, wherein the image processing unit
And extracting a plurality of face images located at different heights within the processed image.

The apparatus of claim 1, wherein the face recognition unit
An area detection unit detecting a plurality of feature areas from the face image;
A normalizing unit for normalizing the face image based on the detected plurality of regions; And
And a matching determination unit for comparing the plurality of points in the normalized face image with a plurality of points of a registered image stored in the database to determine whether to match the plurality of points.

delete

A method of processing a wide-angle image for recognizing a face in an image,
Processing a wide angle image photographed through a wide angle camera module and converting the wide angle image into a planar viewpoint image;
Extracting one or more face images from the transformed planar view image;
Recognizing a face by comparing the extracted one or more face images with registered images stored in a database; And
And tracking movement of the extracted one or more face images,
Wherein the movement information by the step of tracking movement of the extracted one or more face images is stored together with time information.

10. The method according to claim 9,
And converting the spherical coordinates to plane coordinates for the wide-angle image.

10. The method according to claim 9,
And processing the wide-angle image to generate a plurality of planar viewpoint images corresponding to each of a plurality of ROIs.

12. The method of claim 11,
And transmitting at least one planar viewpoint image selected from the plurality of planar viewpoint images to the outside.

12. The method of claim 11,
And storing at least one planar viewpoint image selected from the plurality of planar viewpoint images in a memory.

10. The method of claim 9, wherein the extracting step
And extracting a plurality of face images located at different heights in the converted plane view image.

10. The method according to claim 9, wherein the face recognition step
Detecting a plurality of feature regions from the face image;
Normalizing the face image based on the detected plurality of areas;
And comparing the plurality of points in the normalized face image with a plurality of points of a registered image stored in the database to determine whether or not to match the wide angle image.

delete

16. A recording medium on which a program for causing a computer to execute the method of any one of claims 9 to 15 is recorded.