Augmentation of Image Retrieval using Fractional Coefficients of Hybrid Wavelet Transformed Images with Seven Image Transforms
Hemant B. Kekre, Sudeep D. Thepade, Varun K. Banura, Ankit Khandelwal · 2012
The paper presents novel Content Based Image Retrieval (CBIR) methods using hybrid wavelet transforms generated from the combination of Cosine transform matrix with seven different transforms namely Walsh, Haar, Kekre, Slant, Hartley, Sine (DST) and Cosine (DCT). Here the feature vector size per image is greatly reduced by taking fractional coefficients of the transformed image. The feature vectors are extracted in fifteen different ways from the transformed image. Apart from considering 100 % coefficients of the transformed image, fourteen reduced coefficients sets (as