Video shot boundary detection based on Hilbert and wavelet transform

摘要

In the current work an innovative and robust method of shot boundary detection in a video sequence in the presence of motion and illumination variation is proposed by fusing the advantages of Hilbert transform and DWT so as to get a new feature space and then finding SSIM metric for detecting shot boundaries. The proposed algorithm is tested on TRECVid and publicly available data set. The promising results as compared to the state of the art methods confirms the objective rating of our algorithm.