Video summarization using hidden Markov model

Abstract

This paper proposes a system to analyze and summarize the video shots of the baseball game TV program into fifteen categories. The system consists of three modules: feature extraction, Hidden Markov Model (HMM) training, and video shot categorization.