Title :
Issues in Medical Diagnosis Using Computational Techniques
Author :
Shah, Saurabh ; Parikh, Swapnil
Author_Institution :
Dept. of Comput. Eng., R.K. Univ., Vadodara, India
Abstract :
Early diagnosis of disease may save patient´s life in case of detection of brain tumours and Cancerous tissues, Prediction of Cardiovascular Disease and many more. Computer Aided Diagnosis can assist physician to diagnose disease like cancer early and help consult specialist for further treatment. Paper reviews various techniques presented by researchers for medical diagnosis and their performance issues are discussed. Novel hybrid classifier to classify medical images is proposed. Proposed method first extracts features from images and converts them into transaction database. Then parallel Frequent Pattern (FP) Growth is applied to mine association rules and classified by Decision Tree classifier. Two methods combined together, Parallel FP-Growth association rule mining and Decision Tree classification gives more efficiency and accuracy for proposed system. Various classifiers are compared on real dataset using experiments.
Keywords :
cancer; data mining; decision trees; feature extraction; image classification; medical image processing; tumours; brain tumour detection; cancerous tissue detection; cardiovascular disease prediction; computational technique; computer aided diagnosis; decision tree classifier; disease diagnosis; feature extraction; hybrid classifier; medical diagnosis; medical image classification; parallel FP-Growth association rule mining; parallel frequent pattern growth; patient diagnosis; patient treatment; performance issues; physician assistance; transaction database; Accuracy; Association rules; Biomedical imaging; Classification algorithms; Decision trees; Feature extraction; Association Rule Mining; Classification; FP-tree; Medical Diagnosis; ROI;
Conference_Titel :
Computational Intelligence and Communication Networks (CICN), 2012 Fourth International Conference on
Conference_Location :
Mathura
Print_ISBN :
978-1-4673-2981-1
DOI :
10.1109/CICN.2012.134
