We propose a Bayesian framework for representing and recognizing local image motion in terms of two primitive models: translation and motion discontinuity. Motion discontinuities ...
Boxes are the universal choice for packing, storage, and transportation. In this paper we propose a template-based algorithm for recognition of box-like objects, which is invarian...
The phenomenal growth of video on the web and the increasing sparseness of meta information associated with it forces us to look for signals from the video content for search/info...
Ming Zhao 0003, Jay Yagnik, Hartwig Adam, David Ba...
We present a novel pedestrian detection system based on probabilistic component assembly. A part-based model is proposed which uses three parts consisting of head-shoulder, torso a...
Martin Rapus, Stefan Munder, Gregory Baratoff, Joa...
We present a novel approach to edge detection in bar code signals using a hidden Markov model (HMM). We also present an algorithm for selection of an optimal filter scale used in ...