An Efficient Video Summarization Technique Using Texture Feature and Spectral Clustering

In this paper, a novel approach for video summarization combining spectral clustering and texture feature extraction is presented. Local Optimal Oriented Pattern (LOOP) textural feature extraction technique is explored for each frame of the video in our method for extracting features. Then, a clustering algorithm based on spectral clustering technique is applied to obtain an informative and an efficient video summary by extracting interesting key-frames. We evaluate our an efficient video summarization technique on one standard benchmark SumMe dataset and result clearly illustrates the preeminence of the suggested model in comparison to some of state-of-the-art methods. .