TY - GEN
T1 - Low memory cost block-based belief propagation for stereo correspondence
AU - Tseng, Yu Cheng
AU - Chang, Nelson
AU - Chang, Tian-Sheuan
PY - 2007/12/1
Y1 - 2007/12/1
N2 - The typical belief propagation has good accuracy for stereo correspondence but suffers from large run-time memory cost. In this paper, we propose a block-based belief propagation algorithm for stereo correspondence that partitions an image into regular blocks for optimization. With independently partitioned blocks, the required memory size could be reduced significantly by 99% with slightly degraded performance with a 32×32 block size when compared to original one. Besides, such blocks are also suitable for parallel hardware implementation. Experimental results using Middlebury stereo test bed demonstrate the performance of the proposed method.
AB - The typical belief propagation has good accuracy for stereo correspondence but suffers from large run-time memory cost. In this paper, we propose a block-based belief propagation algorithm for stereo correspondence that partitions an image into regular blocks for optimization. With independently partitioned blocks, the required memory size could be reduced significantly by 99% with slightly degraded performance with a 32×32 block size when compared to original one. Besides, such blocks are also suitable for parallel hardware implementation. Experimental results using Middlebury stereo test bed demonstrate the performance of the proposed method.
UR - http://www.scopus.com/inward/record.url?scp=46449123461&partnerID=8YFLogxK
U2 - 10.1109/ICME.2007.4284925
DO - 10.1109/ICME.2007.4284925
M3 - Conference contribution
AN - SCOPUS:46449123461
SN - 1424410177
SN - 9781424410170
T3 - Proceedings of the 2007 IEEE International Conference on Multimedia and Expo, ICME 2007
SP - 1415
EP - 1418
BT - Proceedings of the 2007 IEEE International Conference on Multimedia and Expo, ICME 2007
Y2 - 2 July 2007 through 5 July 2007
ER -