RegNet: Self-Regulated Network for Image Classification

By J. Xu et al

Published on Jan. 3, 2021

Read the original document by opening this link in a new tab.

Table of Contents

I. Introduction
II. Related Work
III. Our Model
IV. Experiments

Summary

The paper introduces RegNet, a self-regulated network for image classification. It proposes a regulator module based on Convolutional RNNs to extract complementary features in ResNet architectures. Experimental results show promising performance compared to standard ResNet models. The study addresses the depth issue in ResNets by incorporating ConvRNNs. The proposed RegNet architecture outperforms traditional ResNets in classification tasks on CIFAR datasets. The model demonstrates improved accuracy with minimal additional parameters.