We describe a learning-based method for low-level vision problems--estimating scenes from images. We generate a synthetic world of scenes and their corresponding rendered images, m...
Markov Random Field (MRF) models are a popular tool for vision and image processing. Gaussian MRF models are particularly convenient to work with because they can be implemented u...
Marshall F. Tappen, Ce Liu, Edward H. Adelson, Wil...
Markov random fields (MRFs) are popular and generic probabilistic models of prior knowledge in low-level vision. Yet their generative properties are rarely examined, while applica...
In this a novel supervised learning method is proposed to map low-level visualfeatures to high-level semantic conceptsfor region-based image retrieval. The contributions of thispa...
Wei Jiang, Kap Luk Chan, Mingjing Li, HongJiang Zh...
In this work the Low Level Vision Unit (LLVU) of the Heterogeneous and Reconfigurable Machine for Image Analysis (HERMIA) is described. The LLVU consists of the innovative integra...