Sampling has been successfully used to identify performance optimization opportunities. We would like to apply similar techniques to check program correctness. Unfortunately, samp...
We present a two-layer hierarchical formulation to exploit different levels of contextual information in images for robust classification. Each layer is modeled as a conditional f...
We propose a method for trajectory classification based on trajectory voting in Moving Object Databases (MOD). Trajectory voting is performed based on local trajectory similarity. ...
Costas Panagiotakis, Nikos Pelekis, Ioannis Kopana...
Background subtraction is commonly used to detect foreground objects in video surveillance. Traditional background subtraction methods are usually based on the assumption that the...
Partitioning a video sequence into shots is the first step toward video-content analysis and content-based video browsing and retrieval. A video shot is defined as a series of inte...