Tensor Distribution Regression Based on the 3D Conventional Neural Networks

Dear Editor, This letter presents a novel tensor-distribution-regression model based on 3D conventional neural networks (3D-TDR) with an application to clinical score prediction in aging-related diagnosis. The estimation of scores subjects using brain magnetic resonance imaging (MRI) helps understand the pathological stage dementia. However, is still unsolved due reasons of: 1) Analyzing whole-brain MRI extremely difficult as high-dimensional data contains millions voxels; 2) formulated one-dimensional regression issue current deep-learning-based algorithms, which ignores implicit label information between different levels. Motivated by above discoveries, proposed 3D-TDR innovatively establishes following three-fold ideas: a) incorporating tensor layer (TRL) into network (3D-CNN) enable its extraction more discriminative structural changes from (MR) data; b) adopting distribution learning (LDL) fully utilize correlation among MR images, thus emphasizing diversity subjects' scores; and c) combining TRL LDL for end-to-end deep framework, thereby achieving jointly low-rank feature prediction. Experimental results two real-world datasets typical tasks indicate that outperforms benchmark state-of-the-art models. can achieve significant accuracy gain dementia age