Today's category-level object recognition systems largely focus on fronto-parallel views of objects with characteristic texture patterns. To overcome these limitations, we pr...
Abstract. This paper presents a method for tracking multiple skincolored objects in images acquired by a possibly moving camera. The proposed method encompasses a collection of tec...
Although human action recognition has been the subject of much research in the past, the issue of viewpoint invariance has received scarce attention. In this paper, we present an ...
We propose to incorporate hundreds of pre-trained concept detectors to provide contextual information for improving the performance of multimodal video search. The approach takes ...
We propose a method to verify the result of attacks detected by signature-based network intrusion detection systems using lightweight protocol analysis. The observation is that ne...