This work presents a novel machine learning based approach to solve the problem of
video summarization. We present an algorithm to generate the summaries for videos
known to belong to a particular class of videos. In the present work, we use semisupervised
learning to generate a summary and relevance feedback to refine and
personalize the summary.