Performance analysis of Key Frame Extraction using SIFT and SURF algorithms

Growth of videos in today’s Internet usage is extensive. Different types of videos will be available in the Internet which among them are lecture videos. Students can make use of these videos, so there is a need to develop an automated system to search the required content only, rather than wasting the time in viewing the complete video. This can be developed into automated system, required steps are: Frame Extraction, Feature Extraction and Key Frame Extraction. In order to extract the Key Frames, Scale Invariant feature transform (SIFT) and Speed Up Robust Features (SURF) algorithms are used. Accuracy and Robustness are the two main important measures that are considered for performance analysis of computer vision algorithms. This paper presents the performance analysis of SIFT and SURF algorithms in Key Frame Extraction of lecture videos and results show that SURF takes less time when compared to SIFT. Keywords— Frame Extraction, Feature Extraction, Key Frame Extraction, SIFT, SURF