We present a new method for rendering novel images of flexible 3D objects from a small number of example images in correspondence. The strength of the method is the ability to sy...
Abstract We introduce OCELOT, a prototype system for automatically generating the “gist” of a web page by summarizing it. Although most text summarization research to date has ...
Several methods have been developed for face detection with certain success, however these tend to fail under difficult conditions such as partial occlusions and changes in orienta...
Recently there has been considerable interest in topic models based on the bag-of-features representation of images. The strong independence assumption inherent in the bag-of-feat...
This contribution proposes a compositionality architecture for visual object categorization, i.e., learning and recognizing multiple visual object classes in unsegmented, cluttered...