Concept Discoveryin Image Databases using Orthogonal Polynomial Transformation

The objective of this paper is to propose a new algorithm for classifying the concept present in the image. The training images are segmented with fixed size blocks and features are extracted from it. The features are extracted by using the orthogonal polynomials based transformation. From the feature set vectors, a codebook is generated for each concept class. Then significant class representative vectors are calculated which is used for classifying the concepts. The proposed method gives a better representation of training images. The approach definitely guarantees that the results produced are promising.