Low-quality Video Key Frame Extraction via Affinity Propagation

Abstract

With the rapid development of Internet technology and multimedia information, multimedia data, especially video data, has become one of the most important ways to obtain information in people's daily lives. However, it has become a scientific problem to efficiently and quickly extract the information required by users from massive video data. Many years of research in the field of video retrieval show that scientific and rational use of key frame extraction technology can reduce the amount of data generated during video retrieval, save time for video processing, and improve work efficiency. Traditional key frame extraction technology has high requirements for video resolution. For videos that involve privacy protection under low-quality state, most existing literatures do not have targeted solutions. This paper proposes a new key frame extraction algorithm with affinity propagation for video data in the low-quality visual mask domain.