Detecting Structural Alterations in the Brain using a Cellular Neural Network based Classification of Magnetic Resonance Images

The ability to quantify structural attributes using cellular neural networks (CNN) has been shown for a wide range of objects. We here introduce an application that allows the detection of structural alterations in the human brain. Using a CNN-based classification approach we show that a defined class of abnormalities - the so called hippocampal sclerosis - can be detected in T1-weighted magnetic resonance images. Our findings indicate that CNN may prove valuable for a computer-aided diagnosis and classification of images generated by medical imaging systems