Full DST sectorization for feature vector generation in content based image retrieval

Abstract

We have introduced a innovative idea of sectorization of Full DST transformed components. In this paper we have proposed two different approaches along with augmentation of mean of zeroeth and highest row and column components of Full DST transformed image for feature vector generation. The sectorization is performed on two different planes namely Even plane and Even plane. We have introduced the new performance evaluation parameters i. e. LIRS and LSRR apart from precision and Recall, the traditional methods. Two similarity measures such as sum of absolute difference and Euclidean distance are used and results are compared. The cross over point performance of overall average of precision and recall for both approaches on different sector sizes are compared. The DST transform sectorization is experimented on Even and Even plane components of transformed image with augmentation and without augmentation for the color images. The algorithm proposed here is worked over database of 1055 images spread over 12 different classes. Overall Average precision and recall is calculated for the performance evaluation and comparison of 4, 8, 12 & 16 Full DST sectors. The use of sum of Absolute difference as similarity measure always gives lesser computational complexity better result of retrieval.