Stable local feature detection and representation is a fundamental component of many image registration and object recognition algorithms. Mikolajczyk and Schmid [14] recently eva...
Invariant feature descriptors such as SIFT and GLOH have been demonstrated to be very robust for image matching and object recognition. However, such descriptors are typically of ...
In the past, quantized local descriptors have been shown to be a good base for the representation of images, that can be applied to a wide range of tasks. However, current approac...
Abstract. This paper introduces a new framework for image classification using local visual descriptors. The pipeline first performs a nonlinear feature transformation on descripto...
Extraction of stable local invariant features is very important in many computer vision applications, such as image matching, object recognition and image retrieval. Most existing...