Digital images are now the basis of visual information in medical applications. The advent of radiology which employs imaging for diagnosis generates great amount of images. Automatic retrieval of images based on features like color, shape and texture is termed Content Based Image Retrieval. The increasing dependence of modern medicine on diagnostic techniques such as radiology, computerized tomography has resulted in a sudden increase in the number and significance of medical images. Content Based Image Retrieval techniques are being extensively used to aid diagnosis by comparing with similar past cases and improvising Computer Aided Diagnosis. In this paper, it is proposed to extract features in the frequency domain using Walsh Hadamard transform and use FP-Growth association rule mining to extract features based on confidence. The extracted features are classified using Naïve Bayes and CART algorithms and the proposed method’s classification accuracy is evaluated. Experimental results show that classification accuracy for Naïve Bayes is 100 and 96.8 for CART on application of proposed method.
BHUVANESWARI, C.; ARUNA, P.; and LOGANATHAN, D.
"FEATURE SELECTION USING ASSOCIATION RULES FOR CBIR AND COMPUTER AIDED MEDICAL DIAGNOSTIC,"
International Journal of Computer and Communication Technology: Vol. 6:
4, Article 1.
Available at: https://www.interscience.in/ijcct/vol6/iss4/1