Query By Image Content using Discrete Cosine Transform

The need for Query By Image Content (QBIC) peaked up due to the increase in the size of image database. The proposed method chooses appropriate feature extraction methods to extract the features, shape and color to enhance the retrieval efficiency and accuracy. It employs SRM (Statistical Region Merging) algorithm for segmentation and uses DCT (Discrete Cosine Transform) on the segmented image to obtain the shape feature vector. Color feature is obtained by considering the RGB components in the image. The processed feature vectors are collected in the feature database which is then compared with the query image’s feature vector. When the difference matches a specific threshold, the most similar images are retrieved automatically. KeywordsQBIC, SRM, DCT, RGB.