Directional Multiscale Feature Extraction for Biomedical Image Indexing and Retrieval Using Contourlet Transform

Abstract

Multiscale and multidirectional are desirable properties for image decomposition. This paper presents directional multiscale feature extraction for biomedical image indexing and retrieval using contourlet transform. The contourlet transform decomposed image at multiscale in directional sub-bands. The feature vector of each contourlet sub-band is calculated by computing the directional energies of the extracted coefficients of respective sub-band. The final feature vector is constructed by concatenating feature vectors of all contourlet sub-bands. The similarity between query feature vector and feature vector of database is calculated using Manhattan distance. The feature extraction time and feature vector length are reduced significantly using proposed method. The effectiveness of the proposed method is evaluated by conducting the experiments on two well-known biomedical databases: Open access series of imaging studies (OASIS) MRI and NEMA-CT. The average retrieval precision (ARP) and average retrieval rate (ARR) are used to measure the performance of the proposed method. The experimental results show that the proposed method outperforms well-known existing methods.