To be reversible as a data hiding method, the original content and hidden message should be completely recovered. One important objective of this approach is to achieve high embed...
Hee Joon Hwang, Hyoung Joong Kim, Vasiliy Sachnev,...
Many applications in surveillance, monitoring, scientific discovery, and data cleaning require the identification of anomalies. Although many methods have been developed to iden...
We introduce a novel data-driven mean-shift belief propagation
(DDMSBP) method for non-Gaussian MRFs, which
often arise in computer vision applications. With the aid
of scale sp...
Abstract. The mean-shift algorithm is an efficient technique for tracking 2D blobs through an image. Although it is important to adapt the mean-shift kernel to handle changes in il...
The conventional area-based stereo matching algorithm suffers from two problems, the windowing problem and computational cost. Multiple scale analysis has long been adopted in vis...