Abstract. We describe a new approach for estimating the posterior probability of tissue labels. Conventional likelihood models are combined with a curve length prior on boundaries,...
While great strides have been made in detecting and localizing specific objects in natural images, the bottom-up segmentation of unknown, generic objects remains a difficult chall...
Occlusion reasoning, necessary for tasks such as navigation and object search, is an important aspect of everyday life and a fundamental problem in computer vision. We believe tha...
Derek Hoiem, Andrew N. Stein, Alexei A. Efros, Mar...
A qualitative image description grammar with automatic image fitting and object modelling algorithms is presented. The grammar is based on assigning a square sub-region of an imag...
Abstract. We propose a generative approach to the problem of labeling images containing configurations of objects from multiple classes. The main building blocks are dense statisti...