Correspondence Between Spectral Matting and Network Modularity

• Activity Number: 432
• Type: Contributed
• Date/Time: Tuesday, August 6, 2013 : 2:00 PM to 3:50 PM
• Sponsor: International Chinese Statistical Association
• Author(s): Henry Horng-Shing Lu*+ and Hung-Hui Juan and Tung-Yu Wu
• Companies: National Chiao Tung Univ and National Chiao-Tung University and National Chiao-Tung University
• Keywords: Matting ; Spectral Segmentation ; Spectral matting ; Network modularity

Abstract

Spectral matting is a useful technique for resolving the image matting problem. The major issue of spectral matting is to determine the number of matting components, which has significant impacts on the matting performance. In this paper, we propose an improved framework based on spectral matting in order to resolve this issue. First, we transform the spectral matting problem to a network community detection problem, in which the modularity measure is defined to identify how well a community structure is for a given network. To do this, we find the correspondence between the spectral matting and network modularity optimization. Therefore, the modularity optimization is adopted to obtain the appropriate hard segmentation result that can be used as the initial guess of soft matting components. The number of matting components can be determined automatically because the improved framework will search possible image components by iteratively dividing image subgraphs. Furthermore, we also provide an efficient computation approach for the execution of the proposed framework.