Video summarization based on shot boundary detection with penalized contrasts

In this paper, we propose a novel technique for shot boundary detection and video summarization that is based on change point detection with penalized contrasts. The proposed method extracts a proper time series from the video, calculating the mean level of the hue component of consecutive video frames in the Hue Saturation Value color space. Change point detection is applied to the time series, estimating the number of shot transitions and their location. The posterior distribution of the change point sequence is defined, that splits the video into homogenous temporal segments. A representative frame is selected in each segment and similar or meaningless keyframes are deleted from the summary. The resulting summary is of comparable quality to that of the state of the art techniques.