Video Abstraction in H.264/AVC Compressed Domain

Video abstraction allows searching, browsing, and evaluating videos only by accessing their useful contents. Most of the studies performed have used pixel domain, which requires the decoding process and needs more time and process than the compressed domain video abstraction. In this paper, we present a new video abstraction method in the H.264/AVC compressed domain, AVAIF. This method is based upon the normalized histogram of the extracted I-frame prediction modes in the H.264 standard. The frames’ similarities are calculated by intersecting their I-frame prediction modes’ histogram. Moreover, the fuzzy c-means clustering is employed to categorize similar frames and extract key frames. The results obtained show that the proposed method achieves, on average, 85% accuracy and 22% error rate in compressed domain video abstraction, which is higher than the other tested methods in the pixel domain. Moreover, on average, it generates video key frames that are closer to human summaries, and it shows robustness to coding parameters