Our paper “Improving 3D convolutional neural network comprehensibility via interactive visualization of relevance maps: evaluation in Alzheimer's disease” has been accepted by the renowned journal Alzheimer’s Research & Therapy. It describes the extensive evaluation of neural network models to detect Alzheimer’s disease and approaches to generate relevance maps indicating the major contributing brain areas driving the model’s decision. The interactive visualization software accompanying the paper can be accessed here.
Paper accepted by Alzheimer’s Research & Therapy
Zurück zu allen Meldungen