Dementia Detection in MRI Using CNN Classifier

It is essential to emphasise the importance of detecting dementia as early as possible, as this cognitive disorder exhibits signs that increasingly limit the patient. Consequently, early detection aids in improving patient management and controlling disease progression. With the assistance of modern medical imaging techniques, particularly MRI, studying the structural and functional changes in the brain associated with dementia has become feasible. In this work, we focus on the classification of MRI images. We aim to explore the potential of a broad data set of magnetic resonance imaging regarding brain scans to identify markers of cognitive decline, which are often subtle and not easily recognised by clinicians. We expand the range of deep learning applications by implementing algorithms within the medical domain, specifically in tasks involving MRI images. These algorithms are believed to effectively distinguish normal ageing traits from the morphological changes associated with dementia. CNNs enable preferred feature representation in dense layers, minimising the memory consumption typically required for fully connected neural networks. Consequently, CNNs are widely and successfully employed for various applications involving images. Thus, the cognitive impairment of such patients can be diagnosed automatically without the need for labour-intensive and time-consuming preprocessing or manual feature engineering.