Machine learning approach for automatic brain tumour detection using patch-based feature extraction and classification. (27th July 2022)

• Record Type: Journal Article
• Authors: Kalaiselvi, T.; Kumarashankar, P.; Sriramakrishnan, P.
• Is Part Of: International journal of biomedical engineering and technology. Volume 39:Number 4(2022)
• Page Start: 396
• Page End: 411
• Publication Date: 2022-07-27
• Subjects: tumour detection -- feature extraction -- feature blocks -- brain tumour -- BraTS dataset; Biomedical engineering -- Periodicals; 610.28
• Languages: English
• ISSNs: 1752-6418

Abstract

Manual selection of tumorous slices from MRI volume is a time expensive process. In the proposed work, we have developed an automatic method for tumorous slice classification from MRI head volume. The proposed method is named as patch-based classification (PBC). PBC uses 8 × 8 patch-based feature extraction and SVM-based classification. Optimal features selection has been done by infinite feature selection (IFS) method. The optimal features are extracted from patches and used to train the SVM classifier to construct the tumour prediction model (TPM). TPM helps to identify the tumorous patches in the testing phase. Accuracy is measured using sensitivity, specificity and accuracy, error rates are calculated using false alarm (FA) and missed alarm (MA). Training and testing patches are chosen from whole brain atlas (WBA) and BraTS datasets. Results of PBC are compared with several existing methods. PBC will be more helpful to physicians who currently undergo a tedious and time consuming process of manually selecting the tumorous slices from the MRI volumes.