Figure 7
From: A spectrum of sharing: maximization of information content for brain imaging data
Impact of depth of model on classification accuracy in brain imaging data. As the depth of the learner increases (from left to right) the discriminative power of the learnt features increases as well. Notably, the subjects that were held out are also well discriminated, meaning that deep learning generalizes to unseen data. The mapping facilitates analysis of large datasets by displaying complete data in a single figure in a way that highlight data regularities [61].