Novel Image Representation and Description Technique using Density Histogram of Feature Points

This paper introduces novel object shape representation using Density Histogram of Feature Points (DHFP). We use silhouette images where the image region ξ consists of only those pixels that correspond to points on the object and have a value one (1) indicating “on” pixels. We count the number of on pixels in a rectangle boundary around the centroid, in the event that there are no “on” pixels in a rectangle boundary then the value is zero and the rectangle boundaries that are outside the grid are represented by a dummy number. A similarity measure is used to calculate the probability of two image objects being similar. Depending on the value of the probability then a dissimilar can be calculated. This method showed improved retrieval rate due its selective way of calculating dissimilarity of object shapes. Analytic analysis was done to justify our method, experiments were conducted and we tabulated the results.