영상의 특징점과 비용함수를 고려한 스테레오 정합개선 = Improve Stereo Matching by considering the Characteristic Points of the Image and the Cost Function

다국어 초록 (Multilingual Abstract)

This thesis proposes an adaptive variable-sized matching window method using the characteristic points of the image and a method to increase the reliability of the cross-consistency check to raise the correctness of the final disparity image. The proposed adaptive variable-sized window method segments the image with the color information, finds the characteristic points in each segmented image, and varies the size of the matching window according to the existence of the characteristic points inside the window. Also the proposed cross-consistency check method processes the two cases with the cost values corresponding to the best disparity and the second-best disparity: when the cost values themselves are too large and when the difference between the two cost values are too small. The two proposed methods were experimented with the four test images provided by the Middleburry site. As the results from the experiments, the proposed adaptive variable-sized matching window method decreased up to 18.2% of error ratio and the proposed cross-consistency check method increased up to 7.4% of reliability.

국문 초록 (Abstract)

본 논문에서는 최종 변이영상의 정확도를 높이기 위해 영상의 특징점을 이용한 적응적 가변 정합창 방법과 교차 일치성 검사의 신뢰도를 높이는 방법을 제안한다. 제안한 적응적 가변 정합창 방법은 색상정보를 이용하여 영상을 분할하고 분할된 각 영상의 특징점을 찾아 그 특징점들의 유무에 따라 정합창의 크기를 적응적으로 가변시키는 방법이다. 또한 제안한 교차 일치성 검사 방법은 최적의 변이와 차상위 최적의 변이에 대한 비용함수 값들을 비교하여 비용하수 값 자체가 너무 크거나 두 비용함수의 차이가 너무 적은 경우를 찾아내어 처리하는 방법이다. 제안한 두 방법에 대한 Middleburry에서 제공한 네 가지 실험영상을 대상으로 실험한 결과 적응적 가변 정합창 방법은 최대 18.2%의 오차율을 감소시켰다. 또한 제안한 교차 일치성 검사는 최대 7.4%의 신뢰도를 향상시킨 것으로 나타났다.

참고문헌 (Reference)

1 "http://vision.middlebury.edu/stereo"

2 P. Seitz, "Using local orientation information as image primitive for robust object recognition" 1199 : 1630-1639, 1989

3 Jian Sun, "Stereo matching using belief propagation" 27 (27): 787-800, 2003

4 O. Veksler, "Stereo Correspondence with Compact Windows via Minimum Ratio Cycle" 24 : 1654-1660, 2002

5 N. Kim, "Photo-realistic Interactive 3D Virtual Environment Generation Using Multiview Video" 245-254, 2001

6 I. Gallo, "Neural disparity computation for dense two-frame stereo correspondence" 29 (29): 673-687, 2008

7 Z. Gu, "Local stereo matching with adaptive support-weight, rank transform and disparity calibration" 29 : 1230-1235, 2008

8 "ISO/IEC MPEG & ITU-T VCEG, “Multiview video plus depth (MVD) format for advanced 3D video systems,” JVT-W100"

9 A. Fusiello, "Experiments with a new area-based stereo algorithm" 1997

10 J. Schimidt, "Dense Disparity Maps in Real-Time with an Application to Augmented Reality" 225-230, 2002

11 V. Kolmogorov, "Computing visual correspondence with occlusions via graph cuts" 2 : 508-515, 2001

12 H. Saito, "Appearance-based Virtual View Generation of Temporally-Varying Events from Multi-Camera Images in the 3D Room" 5 (5): 303-316, 2003

13 Kuk-Jin Yoon, "Adaptive Support-Weight Approach for Correspondence Search" 2006

14 T. Kanade, "A stereo matching algorithm with an adaptive window: theory and experiment" 16 : 920-932, 1994

15 I. J. Cox, "A maximum likelihood stereo algorithm" 63 (63): 542-567, 1996

16 T. Kanade, "A Stereo Matching Algorithm with an Adaptive Window: Theory and Experiments" 16 : 920-932, 1994

17 Peter N. Belhumeur, "A Bayesian approach to binocular stereopsis" 19 (19): 237-260, 1996