RGB-D 정보 및 거리변환을 이용한 보행자 검출

다국어 초록 (Multilingual Abstract)

According to the development of depth sensing devices and depth estimation technology, depth information becomes more important for object detection in computer vision. In terms of recognition rate, pedestrian detection methods have been improved more accurately. However, the methods makes slower detection time. So, many researches have overcome this problem by using GPU. Here, we propose a real-time pedestrian detection algorithm that does not rely on GPU. First, the depth-weighted distance map is used for detecting expected human regions. Next, human detection is performed on the regions. The performance for the proposed approach is evaluated and compared with the previous methods. We show that proposed method can detect human about 7 times faster than conventional ones.

목차 (Table of Contents)

• Abstract
• 1. 서론
• 2. 예상 영역을 이용한 보행자 검출 알고리즘
• 3. 실험 및 결과
• 4. 결론
• References

참고문헌 (Reference)

1 김재도, "적분영상 기반 특징 정보 예측을 통한 고속 보행자 검출" 한국전기전자학회 17 (17): 469-477, 2013

2 박병주, "실시간 물체 검출을 위한 고효율 Viola-Jones 검출 프레임워크" 한국전기전자학회 18 (18): 1-7, 2014

3 R. Benenson, "Ten Years of Pedestrian Detection, What Have We Learned" 2014

4 Jun Liu, "Real-time Human Detection and Tracking in Complex Environments using Single RGB-D Camera" 3088-3092, 2013

5 P. Viola, "Rapid Object Detection using a Boosted Cascade of Simple Features" 1 : 511-518, 2001

6 L. Spinello, "People Detection in RGB-D Data" 3838-3843, 2011

7 P. Dollar, "Intergral Channel features" 2009

8 A. Satpathy, "Human detection by quadratic classification on subspcae of extended histogram of gradients" 3 (3): 287-297, 2014

9 N. Dalal, "Histograms of Oriented Gradients for Human Detection" 1 : 886-893, 2005

10 R. Kimmel, "Distance maps and Weighted Distance Transforms" 6 : 223-233, 1996

11 G. Borgefors, "Distance Transformations in Digital Images" 34 : 344-371, 1986

12 G. Borgefors, "Distance Transformations in Arbitrary Dimensions" 27 : 321-345, 1984

13 Yujie Shen, "A Novel Human Detection Approach Based on Depth Map via Kinect" 535-541, 2013