Constrained Linear Data-feature Mapping for Image Classification
From MaRDI portal
Publication:6329760
arXiv1911.10428MaRDI QIDQ6329760
Author name not available (Why is that?)
Publication date: 23 November 2019
Abstract: In this paper, we propose a constrained linear data-feature mapping model as an interpretable mathematical model for image classification using convolutional neural network (CNN) such as the ResNet. From this viewpoint, we establish the detailed connections in a technical level between the traditional iterative schemes for constrained linear system and the architecture for the basic blocks of ResNet. Under these connections, we propose some natural modifications of ResNet type models which will have less parameters but still maintain almost the same accuracy as these corresponding original models. Some numerical experiments are shown to demonstrate the validity of this constrained learning data-feature mapping assumption.
Has companion code repository: https://github.com/xuteam/mgnet_code
This page was built for publication: Constrained Linear Data-feature Mapping for Image Classification
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6329760)
