SPATIAL PYRAMID KERNEL FOR IMAGE CLASSIFICATION

This paper presents a method for recognizing toy pieces categories based on the global features that describe color and illumination properties, and by using the statistical learning paradigm. This technique works by partitioning the image into increasingly finer sub-regions and computing a weighted histogram intersection in this space. First, we show the mathematical properties of it to be used as a kernel function for Support Vector Machines (SVMs). Second, we describe the implementation and give examples of how these SVMs, equipped with such a kernel, can achieve very promising results on image classification.