Pixel-Wise Histograms for Visual Segment Description and Applications

Automatic Key-Frame Extraction From Broadcast Soccer Videos

“…A key-frame is a representative image of a shot and its extraction is an important tool for the process of video semantic analysis (Ciocca and Schettini, 2005;Doulamis et al, 2000;Dufaux, 2000;Wolf, 1996). In general, it is frequently used for video visualization (Arman et al, 1994;Komlodi and Marchionini, 1998;Tse et al, 1998;Zhong et al, 1996) and recovery (Arman et al, 1994;Liu et al, 2003;Sze et al, 2005;Pardo, 2006), or even as a tool for scenes detection (Xiong et al, 1997).…”

“…In such case, the obtained key-frame is a composition of different frames which yields a key-frame representation that can be useful for video retrieval but not for video browsing and navigation, since this synthetic key-frame is not necessarily included in the set of the original frames belonging to the corresponding shot. (Pardo, 2006) also considers a video segment or a group of frames to compute the pixel-wise histogram which, again, can be useful for video retrieval but not for video browsing and navigation.…”

Video Analysis Via Nonlinear Dimensionality Reduction

“…Some of these methods can be rooted to image analysis method that use all the pixels and visualize the image as a point in a high dimensional space [7]. Other methods for video analysis recently proposed in the literature use pixel-wise histograms to describe video segments [6]. These methods keep a lot of information from the beginning of the processing.…”