Abstract. We exploit the recurrence of structures at different locations, orientations and scales in an image to perform denoising. While previous methods based on "nonlocal f...
Yifei Lou, Paolo Favaro, Stefano Soatto, Andrea L....
Opinion finding is a challenging retrieval task, where it has been shown that it is especially difficult to improve over a strongly performing topic-relevance baseline. In this pa...
Rodrygo L. T. Santos, Ben He, Craig Macdonald, Iad...
There are many computer vision algorithms developed for visual (scene and object) recognition. Some systems focus on involved learning algorithms, some leverage millions of trainin...
A framework is presented for estimating the pose of a camera based on images extracted from a single omnidirectional image of an urban scene, given a 2D map with building outlines...
This paper presents a new approach for multi-view object class detection. Appearance and geometry are treated as separate learning tasks with different training data. Our approach...