계층적 불변 모멘트를 이용한 외곽선 영상기반 ASL 숫자 인식 시스템 개발

다국어 초록 (Multilingual Abstract)

This paper propose the hand shape recognition using the Hierarchical Moment Invariants. First, to reduce the error rate, HSV region color space is used. Detected binary image is opening calculated which containing erosion and dilation to erase the noise. Next, obtain the edge line using Canny edge and then implement the Hierarchical Moment Invariants. Hierarchical layers consist of Orientation layer, Objet external layer and Moment Invariants layer. In the Orientation layer, ellipse tilt detecting algorithm is used to find the orientation of the hand. In the Object external layer, Circularity and Eccentricity is used to find the feature value. Last layer can obtain the 3 feature of the Moment invariants and then all of the features are recognized by BP artificial neural network. In recognizing the 9 ASL, recognition rate shows 97.15% and FPS shows 0.046.

참고문헌 (Reference)

1 주란희, "영역 중심 모멘트와 장선 특징을 이용한 아무르불가사리 다중개체 인식 기법" 한국인터넷방송통신학회 10 (10): 83-88, 2010

2 이동형, "순차 램 기반 누적 신경망을 이용한 수화 인식" 한국인터넷방송통신학회 9 (9): 205-211, 2009

3 박기서, "깊이 정보와 비주얼 영상을 이용한 손동작 인식" 한국정보기술학회 11 (11): 57-65, 2013

4 M. K. Hu, "Visual pattern recognition by moment invariants" 8 : 179-187, 1962

5 Johannes Kilian, "Simple Image Analysis By Moments Version 2.0"

6 J. Flusser, "On the independence of rotation moment invariants" 33 (33): 1405-1410, 2000

7 C. S. Oh, "Neuro computer" 194-227, 2000

8 Guang-Bin Huang, "Learning Capability and Storage Capacity of Two-Hidden-Layer Feedforward Networks" IEEE 274-281, 2003

9 M. R. Teague, "Image analysis via the general theory of moments" 70 (70): 920-930, 1980

10 F. O. Araujo T., "Detecting hands, fingers and blobs for multi-touch display applications" IEEE 237-243, 2009

11 Bao, Paul, "Canny edge detection enhancement by scale multiplication" IEEE 1485-1490, 2005

12 P. J. Werbos, "Backpropagation through time:what it does and how to do it" 1990

13 Jung-Hyun Park, "A Research on SO-Moment Invariants Based on Edge Detection for Real-time Hand Shape Recognition" 38-41, 2014

14 Omer Rashid, "A Framework for the Integration of Gesture and Posture Recognition Using HMM and SVM" IEEE 4 : 572-577, 2009