作者: Kadir Peker , Faisal Bashir
DOI:
关键词:
摘要: A method summarizes a video including sequence of frames. The is partitioned into segments frames, and faces are detected in the frames segments. Features extracted. For each segment faces, representative frame based on features selected. possible pair distances determined faces. arranged matrix. Spectral clustering applied to matrix determine an optimal number clusters. Then, can be summarized according