Biomedical images and captions are one of the major sources of information in online biomedical publications. They often contain the most important results to be reported, and pro...
Xin Chen, Caimei Lu, Yuan An, Palakorn Achananupar...
This paper presents a generative model based approach to man-made structure detection in 2D natural images. The proposed approach uses a causal multiscale random field suggested i...
Parameterized Appearance Models (PAMs) (e.g. eigentracking, active appearance models, morphable models) use Principal Component Analysis (PCA) to model the shape and appearance of...
This paper introduces a simple and efficient representation for natural images. We partition an image into blocks and treat the blocks as vectors in a high-dimensional space. We t...
We derive a probabilistic multi-scale model for contour completion based on image statistics. The boundaries of human segmented images are used as "ground truth". A prob...