Scene recognition via scene template learning

This paper concerns the problem of scene recognition in an image possibly containing a single or multiple scenes. The first half of the paper will concentrate on a newly built model, called scene template, and come up with a set of well developed methodologies to learn the model. The second half of the paper will introduce the technique used to recognize scenes, assuming an ideal model is given, or once the model is well trained. Future Distribution Permission The author(s) of this report give permission for this document to be distributed to Stanfordaffiliated students taking future courses.