This paper presents an edge-directed super-resolution algorithm for gray level document images without using any training set. This technique creates an image with smooth regions ...
We seek to both detect and segment objects in images. To exploit both local image data as well as contextual information, we introduce Boosted Random Fields (BRFs), which use boos...
Antonio Torralba, Kevin P. Murphy, William T. Free...
3D Bayesian regularization applied to diffusion tensor MRI is presented here. The approach uses Markov Random Field ideas and is based upon the definition of a 3D neighborhood syst...
Extracting titles from a PDFs full text is an important task in information retrieval to identify PDFs. Existing approaches apply complicated and expensive (in terms of calculating...
Abstract. In this paper, we propose an unsupervised color image segmentation scheme using homotopy continuation method and Compound Markov Random Field (CMRF) model. The proposed s...