Classification of CT brain images based on deep learning networks

• A fused CNN architecture achieving classification accuracy rate of 87.62%.• 2D CNN delivers 86.32% precision for the 3 classes of AD, Lesion, Normal.• 2D SIFT and 2D KAZE give accuracy rates of 85.61% and 86.31% respectively.• 3D SIFT and 3D KAZE achieve accuracy rates of 85.26% and 83.15% respectively.

Source: Computer Methods and Programs in Biomedicine - October 19, 2016 Category: Bioinformatics Authors: Xiaohong W. Gao, Rui Hui, Zengmin Tian Source Type: research