Abstract—This correspondence presents an image compression approach that integrates our proposed parameter-assistant inpainting (PAI) to exploit visual redundancy in color images...
Object detection is challenging partly due to the limited discriminative power of local feature descriptors. We amend this limitation by incorporating spatial constraints among ne...
In this paper, we present a novel method for human action recognition from any arbitrary view image sequence that uses the Cartesian component of optical flow velocity and human ...
This paper presents a cluster validation based document clustering algorithm, which is capable of identifying both important feature words and true model order (cluster number). I...
Most automatic expression analysis systems attempt to recognize a small set of prototypic expressions (e.g. happiness and anger). Such prototypic expressions, however, occur infre...