James Thewlis

  • Citations Per Year
Learn More
Learning automatically the structure of object categories remains an important open problem in computer vision. In this paper, we propose a novel unsupervised approach that can discover and learn landmarks in object categories, thus characterizing their structure. Our approach is based on factorizing image deformations, as induced by a viewpoint change or(More)
genaue r b e k a n n t sein. Mit se inen Mitarbei ter i l konn t e R. M A ~ K O ~ den Einf luB der T e m p e r a t u r au f die I6n isa t ion der A tome zeigen. Das wieh t igs te Ergebn i s dieser U n t e r s u c h u n g e n dfirfte der Naehweis reiner T e m p e r a t u r a n r e g u n g ~fir s t roms~arke L i ch t b6gen sein. N a c h P, espreehui lg der(More)
One of the key challenges of visual perception is to extract abstract models of 3D objects and object categories from visual measurements, which are affected by complex nuisance factors such as viewpoint, occlusion, motion, and deformations. Starting from the recent idea of viewpoint factorization, we propose a new approach that, given a large number of(More)
As a result of renewed interest in Convolutional Neural Networks (CNNs) for visual recognition following their success in the ILSVRC 2012 challenge (Krizhevsky, Sutskever, and Hinton, 2012), along with the surprising representational ability of features extracted from such networks in a wide range of computer vision problems (Donahue et al., 2013; Razavian(More)
  • 1