KR970002750A - Face Recognition Method in Image Data - Google Patents

Face Recognition Method in Image Data Download PDF

Info

Publication number
KR970002750A
KR970002750A KR1019950016017A KR19950016017A KR970002750A KR 970002750 A KR970002750 A KR 970002750A KR 1019950016017 A KR1019950016017 A KR 1019950016017A KR 19950016017 A KR19950016017 A KR 19950016017A KR 970002750 A KR970002750 A KR 970002750A
Authority
KR
South Korea
Prior art keywords
face recognition
velocity vector
recognition method
face
image
Prior art date
Application number
KR1019950016017A
Other languages
Korean (ko)
Inventor
박규호
김준성
Original Assignee
심상철
한국과학기술원
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 심상철, 한국과학기술원 filed Critical 심상철
Priority to KR1019950016017A priority Critical patent/KR970002750A/en
Publication of KR970002750A publication Critical patent/KR970002750A/en

Links

Landscapes

  • Image Analysis (AREA)

Abstract

본 발명은 영상자료에 있어서의 얼굴인식방법에 관한 것이다. 좀 더 구체적으로, 본 발명은 일련의 얼굴영상으로부터의 움직임 정보를 이용하여 효율적이면서 고속으로 얼굴인식을 수행할 수 있는 영상자료에 있어서의 얼굴인식방법에 관한 것이다. 본 발명에 따른 얼굴인식방법은 일정한 크기의 영역 내에서 얻어진 다수의 영상점의 공간 및 시간적인 변화율을 산출하는 미분과정(10); 전기한 미분과정(10)에 의해 산출된 변화율에 의거하여 가능한 다수의 속도벡터를 산출하고 가장 부합되는 속도벡터장을 결정하는 영상 속도벡터 산출과정(20); 전기 과정(20)에 의해 산출된 속도벡터장의 크기로부터 움직이는 영영과 움직이지 않는 배경을 구분하는 분리 과정(30); 및 전기 과정(30)에 의해 얻어진 분리결과로부터 타원 근사화법을 통하여 얼굴부의 경사정도를 파악하여 회전보정을 행하는 회전보정과정(40)을 포함한다. 본 발명의 얼굴인식방법은 조명이나 배경상태 등에 영향을 받지 않아 오류발생없이 얼굴인식을 수행할 수 있으며, 국부적인 연산에 의해서도 효율적이면서 고속으로 얼굴인식을 수행할 수 있는 등의 효과를 지니고 있다.The present invention relates to a face recognition method in video data. More specifically, the present invention relates to a face recognition method in image data capable of performing face recognition efficiently and at high speed by using motion information from a series of face images. The face recognition method according to the present invention includes a differential process (10) for calculating the spatial and temporal rate of change of a plurality of image points obtained in a region of a constant size; An image velocity vector calculation process (20) for calculating a plurality of velocity vectors based on the change rate calculated by the aforementioned differential process (10) and determining the most suitable velocity vector field; A separation process 30 for distinguishing the moving zero and the non-moving background from the magnitude of the velocity vector field calculated by the electric process 20; And a rotation correction process 40 which performs rotation correction by grasping the inclination degree of the face portion through the elliptic approximation method from the separation result obtained by the electric process 30. The face recognition method of the present invention is capable of performing face recognition without generating an error because it is not influenced by lighting or background state, and has an effect of performing face recognition at high speed even by local operation.

Description

영상자료에 있어서의 얼굴인식방법Face Recognition Method in Image Data

본 내용은 요부공개 건이므로 전문내용을 수록하지 않았음Since this is an open matter, no full text was included.

제1도는 본 발명의 얼굴인식방법에 의한 얼굴인식시 사용되는 얼굴인식 시스템의 블럭구성도이다. 제2도는 본 발명에 따른 얼굴인식방법을 도식적으로 나타낸 흐름도이다.1 is a block diagram of a face recognition system used in face recognition by the face recognition method of the present invention. 2 is a flowchart schematically showing a face recognition method according to the present invention.

Claims (6)

(ⅰ) 일정한 크기의 영역 내에서 얻어진 다수의 영상점의 공간 및 시간적인 변화율을 산출하는 미분과정(10); (ⅱ) 전기한 미분과정(10)에 의해 산출된 변화율에 의거하여 가능한 다수의 속도벡터를 산출하고 가장 부합되는 속도벡터장을 결정하는 영상 속도벡터 산출과정(20); (ⅲ) 전기 과정(20)에 의해 산출된 속도벡터장의 크기로부터 움직이는 영역과 움직이지 않는 배경을 구분하는 분리과정(30); 및, (ⅳ) 전기 과정(30)에 의해 얻어진 분리결과로부터 타원 근사화법을 통하여 얼굴부의 경사정도를 파악하여 회전보정을 행하는 회전보정과정(40)을 포함하는 영상자료에 있어서의 얼굴인식방법.(Iii) a differential process (10) for calculating the spatial and temporal rate of change of the plurality of image points obtained within a region of constant size; (Ii) an image velocity vector calculation process (20) for calculating a plurality of possible velocity vectors based on the change rate calculated by the aforementioned differential process (10) and determining the most suitable velocity vector field; (Iii) a separation process 30 for distinguishing the moving region from the non-moving background from the magnitude of the velocity vector field calculated by the electrical process 20; And (i) a rotation correction process (40) which performs rotational correction by grasping the inclination degree of the face portion through an elliptic approximation method from the separation result obtained by the electrical process (30). 제1항에 있어서, 전기한 회전보정과정(40)에서 확정된 얼굴영상 데이타를 기저장된 얼굴영상 데이타와 비교하는 얼굴인식과정(50)을 추가로 포함하는 것을 특징으로 하는 영상자료에 있어서의 얼굴인식방법.The face in the image data according to claim 1, further comprising a face recognition process (50) for comparing the face image data determined in the rotation correction process (40) described above with previously stored face image data. Recognition method. 제1항에 있어서, 미분과정(10)중 공간적인 변화율은 연속적인 3매의 영상프레임의 x방향 및 y방향에 대한 변화를 계산하고 그 결과를 평균하여 얻어지는 것을 특징으로 하는 영상자료에 있어서의 얼굴인식방법.The method of claim 1, wherein the spatial rate of change in the differential process (10) is obtained by calculating the change in the x direction and the y direction of three consecutive image frames and averaging the results. Face recognition method. 제1항에 있어서, 미분과정(10)중 시간적인 변화율은 연속적인 3매의 영상프레임에 따른 밝기의 변화를 계산하여 얻어지는 것을 특징으로 하는 영상자료에 있어서의 얼굴인식방법.The method of claim 1, wherein the temporal change rate during the differential process (10) is obtained by calculating a change in brightness according to three consecutive image frames. 제1항에 있어서, 전기한 영상 속도벡터 산출과정(20)에서는 전기한 산출과정(20)을 속도벡터장을 구하고자 하는 점의 주위에 대해서도 적용하여 가능한 다수의 속도벡터를 구하고 소정의 기준에 부합하는 속도벡터를 산출하여 원하는 속도벡터장을 얻는 것을 특징으로 하는 영상자료에 있어서의 얼굴인식방법.The method according to claim 1, wherein in the above-described image velocity vector calculation process (20), the aforementioned calculation process (20) is also applied to the periphery of the point for which the velocity vector field is to be obtained. A face recognition method in image data, characterized by obtaining a desired velocity vector field by calculating a corresponding velocity vector. 제2항에 있어서, 얼굴인식방법의 수행시에는 영상화면을 아날로그 영상 전기 신호로 변환하는 비데오 카메라(1)와; 전기 비데오 카메라(1)로부터 출력된 아날로그 영상신호를 디지탈 영상신호로 변환하는 인터페이스부(2)와; 전기 인터페이스부(2)를 통하여 변환된 디지탈 영상신호를 처리하여 얼굴부를 추출하고 형태학적으로 분석하여 분석된 형태학적 데이타를 기저장된 기지의 얼굴영상 데이타와 비교하여 신원을 확인하는 메인컴퓨터(3)와; 전기한 얼굴영상 및 신원에 관한 정보를 시각적으로 표시하는 모니터(4); 및 이를 출력하기 위한 프린터(5)로 구성된 얼굴인식 시스템을 사용하는 것을 특징으로 하는 영상자료에 있어서의 얼굴인식방법.A video camera (1) according to claim 2, further comprising: a video camera (1) for converting a video screen into an analog video electrical signal when performing a face recognition method; An interface unit 2 for converting an analog video signal output from the electric video camera 1 into a digital video signal; The main computer (3) for processing the digital image signal converted through the electrical interface unit (2), extracting the face part, and morphologically analyzing and comparing the analyzed morphological data with previously stored facial image data to confirm the identity. Wow; A monitor 4 for visually displaying the facial image and the information on the identity described above; And a face recognition system comprising a printer (5) for outputting the face recognition method. ※ 참고사항 : 최초출원 내용에 의하여 공개하는 것임.※ Note: The disclosure is based on the initial application.
KR1019950016017A 1995-06-16 1995-06-16 Face Recognition Method in Image Data KR970002750A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
KR1019950016017A KR970002750A (en) 1995-06-16 1995-06-16 Face Recognition Method in Image Data

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
KR1019950016017A KR970002750A (en) 1995-06-16 1995-06-16 Face Recognition Method in Image Data

Publications (1)

Publication Number Publication Date
KR970002750A true KR970002750A (en) 1997-01-28

Family

ID=66524383

Family Applications (1)

Application Number Title Priority Date Filing Date
KR1019950016017A KR970002750A (en) 1995-06-16 1995-06-16 Face Recognition Method in Image Data

Country Status (1)

Country Link
KR (1) KR970002750A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100338473B1 (en) * 1999-07-02 2002-05-30 조양호 Face detection method using multi-dimensional neural network and device for the same
KR100900284B1 (en) * 2002-05-31 2009-05-29 엘지전자 주식회사 Mobile terminal and Method for searching a phone number in thereof
KR20120021821A (en) * 2010-08-18 2012-03-09 주식회사 바이오스페이스 Body composition analyzer had at least one camera

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100338473B1 (en) * 1999-07-02 2002-05-30 조양호 Face detection method using multi-dimensional neural network and device for the same
KR100900284B1 (en) * 2002-05-31 2009-05-29 엘지전자 주식회사 Mobile terminal and Method for searching a phone number in thereof
KR20120021821A (en) * 2010-08-18 2012-03-09 주식회사 바이오스페이스 Body composition analyzer had at least one camera

Similar Documents

Publication Publication Date Title
US5274453A (en) Image processing system
Yacoob Computing spatio-temporal representations of human faces
KR100423379B1 (en) Key signal generating device, picture producing device, key signal generating method, and picture producing method
US20040105493A1 (en) Image processing apparatus and method, and image pickup apparatus
KR960039982A (en) Moving object segmentation of motion image and its motion estimation method
WO2002093488A3 (en) Segmentation unit for and method of determining a second segment and image processing apparatus
JPH0916783A (en) Apparatus and method for outline image detection/thinning ofobject
US20070274402A1 (en) Application of short term and long term background scene dynamics in motion detection
JP2002190016A (en) Signal processor, method and recording medium
JP2002190028A (en) Device and method for processing signal, and recording medium
KR100846248B1 (en) Image processor
Karabasi et al. A model for real-time recognition and textual representation of malaysian sign language through image processing
KR960028564A (en) Motion vector selection method and image processing device performing the method
KR970002750A (en) Face Recognition Method in Image Data
JP2004530214A (en) Text discrimination method and related apparatus
JP2000134638A (en) Image processor
JPH0991453A (en) Image processing method and its unit
JP2001307104A (en) Object extraction device for moving image
Jain et al. Analyzing a real world scene sequence using fuzziness
Tucker et al. Image stabilization for a camera on a moving platform
JP2001025021A (en) Motion detection method and motion detector
KR940001733A (en) Motion Estimation Method in Image Data Compression and Its Apparatus
KR100265812B1 (en) Apparatus and method for motion segmentation of image sequence
JP2000261741A (en) Still picture extraction device and program storage medium for extracting still picture
JP2645556B2 (en) Digital video display

Legal Events

Date Code Title Description
A201 Request for examination
E902 Notification of reason for refusal
E601 Decision to refuse application