Abstract—In this work, efficient medical image retrieval is made possible by using texture features. The texture features used are discrete wavelet transform and rotated complex wavelet transform. The discrete wavelet transform uses Haar wavelet. The reason is, because of its easy computation. A two dimensional (2-D) Rotated wavelet are designed with complex wavelet filter coefficients, which gives texture information strongly oriented in six different directions (45° apart from complex wavelet transform). This texture image retrieval system is capable of providing retrieval result with high retrieval accuracy and less computational complexity. Proposed a novel approach for texture image retrieval by using two approaches, one by using Discrete Wavelet Transform (DWT) and second by a set of Dual-Tree Rotated Complex Wavelet Filter (DT-RCWF). The information provided by DT-RCWF and DWT were analyzed in detail. Features are obtained by computing the energy on each sub band of the decomposed image. To check the retrieval performance, Texture database is created using images like ultrasound, CT and MRI of 600 images. The algorithm of DWT and RCWF was applied on these databases and their performances were evaluated. The retrieval rate of DWT outperforms RCWF. RCWF shows less retrieval rate in precision and recall values as 60%and 30% comparing while traditional discrete wavelet transform based approach with 80% and 70% while retrieving the medical image. The proposed method also retains comparable levels of computational complexity.

Index Terms—Medical image retrieval, Discrete wavelet transform, Rotated complex wavelet transform, Texture features

Cite: Swarnambiga A, Vasuki S and Anantha Raja A, "MEDICAL IMAGE RETRIEVAL USING TEXTURE FEATURES BY DISCRETE WAVELET TRANSFORM AND ROTATED COMPLEX WAVELET TRANSFORM," International Journal of Electrical and Electronic Engineering & Telecommunications, Vol. 1, No. 1, pp. 17-29, October 2012.