Object representation using SIFT and local contour

This paper proposes a new object representation method based on local sift descriptor and contour. First local feature transform into sparse code based on Locality-constrained Linear Coding (LLC), then local sparse codes are integrated together for the whole object image using spatial pyramid model. The approach can be applied to any object with distinguishable parts in a relatively fixed spatial configuration. Based on this representation, a learning algorithm is used to automatically learn to detect instances of the object class in new images. It is evaluated here on difficult sets of real-world images containing cars, and is seen to successfully detect objects in varying conditions amidst background clutter and mild occlusion. In addition this paper also uses some prior information to construct a real-time car detection system. © 2011 Binary Information Press December, 2011.