In human physiology, the brain plays a significant role as the control center of all regulatory processes. Any abnormality in the brain could lead to various physiological and psychological problems and, thus, demands early detection and treatment. Accurate identification of a brain condition is vital during treatment planning. Hence, medical imaging combined with Artificial Intelligence (AI) schemes is widely employed in hospitals to achieve better detection accuracy. Specifically, a brain hemorrhage is a medical emergency that needs immediate treatment to reduce its impact. Therefore, this research aimed to develop and implement a Lightweight Deep Learning (LDL) procedure to classify brain Computed Tomography (CT) slices into healthy/hemorrhagic classes. The various stages of this scheme involve: (i) image collection, resizing, and preprocessing; (ii) LDL feature extraction; and (iii) binary classification with 3-fold cross-validation. In this work, the CT slices were initially preprocessed with a threshold filter and then considered to verify the performance of the proposed scheme based on individual and dual features. The experimental outcome of this study con-firms that the dual features help to achieve a detection accuracy of >96% with the Support Vector Machine (SVM) classifier.
Discussion(0)
No comments yet. Be the first to comment.