Title :
A dynamic programming approach for appearance-based recognition of environments
Author :
Chen, Chu-Song ; Chen, Jiun-Hung ; Hsieh, Wen-Teng
Author_Institution :
Inst. of Inf. Sci., Acad. Sinica, Taipei, Taiwan
Abstract :
A systematic method for recognizing the contents of a video through the help of some panoramas recorded in a database is proposed in this paper. A panorama inherently records the appearances of an omni-directional scene from its center point to arbitrary viewing directions, and thus can serve as a compact representation of an environment. The associated recognition task is formulated as a shortest-path search problem, and a dynamic-programming technique is used to solve it. Experimental results show that our method can recognize a video effectively.
Keywords :
dynamic programming; image matching; image recognition; image representation; image segmentation; image sequences; search problems; video signal processing; appearance-based environment recognition; arbitrary viewing directions; candidate-selection stage; compact environment representation; dynamic-programming technique; graph-construction stage; image-sequence-based method; matching reliability; omni-directional scene; panorama; path-searching stage; shortest-path search problem; systematic method; video contents recognition; view-based object representation; Dynamic programming; Image databases; Image recognition; Image segmentation; Image sequences; Information science; Layout; Neural networks; Object detection; Visual databases;
Conference_Titel :
Pattern Recognition, 2002. Proceedings. 16th International Conference on
Print_ISBN :
0-7695-1695-X
DOI :
10.1109/ICPR.2002.1048448
