Pei Ji-hong, Ye Pei-ling, Xie Wei-xin. Color Optical Flow Estimation Based on Quaternion Homogeneity Differential[J]. Journal of Electronics & Information Technology, 2009, 31(11): 2614-2619. doi: 10.3724/SP.J.1146.2008.01469
Citation:
Pei Ji-hong, Ye Pei-ling, Xie Wei-xin. Color Optical Flow Estimation Based on Quaternion Homogeneity Differential[J]. Journal of Electronics & Information Technology, 2009, 31(11): 2614-2619. doi: 10.3724/SP.J.1146.2008.01469
Pei Ji-hong, Ye Pei-ling, Xie Wei-xin. Color Optical Flow Estimation Based on Quaternion Homogeneity Differential[J]. Journal of Electronics & Information Technology, 2009, 31(11): 2614-2619. doi: 10.3724/SP.J.1146.2008.01469
Citation:
Pei Ji-hong, Ye Pei-ling, Xie Wei-xin. Color Optical Flow Estimation Based on Quaternion Homogeneity Differential[J]. Journal of Electronics & Information Technology, 2009, 31(11): 2614-2619. doi: 10.3724/SP.J.1146.2008.01469
Due to representing 3-dimensional color vectors as quaternion hypercomplex numbers, a color image sequences is mapped as a quaternion hypercomplex function in 3-dimensional spatial-temporal space. By underlying hypothesis of Color Constant Model (CCM), a quaternion Optical Flow Equation (OFE) which is used to describe motion in color image sequences is developed. In order to solve this quaternion OFE efficiently, a concept of quaternion homogeneity differential is introduced, thus that a new Quaternion Homogeneity CCM (QHCCM) OFE is built. The mathematic form of the QHCCM OFE is similar with the Brightness Constant Model (BCM) OFE. Because it contains not only luminance information but also chrominance information, the QHCCM OPE can capture motions in color video more accurate than that of the BCM-based method. Finally, some compared experiments are given to show the proposed methods effectiveness.