Mean shift is a popular method to segment images and videos. Pixels are represented by feature points, and the segmentation is driven by the point density in feature space. In thi...
Many applications in surveillance, monitoring, scientific discovery, and data cleaning require the identification of anomalies. Although many methods have been developed to iden...
Image perception in underwater environment is a difficult task for a human operator, and data segmentation becomes a crucial step toward an higher level interpretation and recogni...
Umberto Castellani, Marco Cristani, Vittorio Murin...
This work deals with the problem of backgroundforeground segmentation in video scenes. We propose an approach that makes use of feature extraction and matching, robust estimation,...
The mean shift (MS) algorithm is sensitive to local peaks. In this paper, we show both empirically and analytically that when using sample data, the reconstructed PDF may have fals...