Ekconv: compressing convolutional neural networks with evolutionary kernel convolution

D Yang, Z Chen, Y Sun, Q He, S Ye… - Journal of Physics …, 2023 - iopscience.iop.org
Convolutional neural networks (CNNs) have achieved tremendous success in visual
recognition tasks but mainly rely on massive learnable parameters. To solve this problem …

Circular FC: Fast Fourier Transform Meets Fully Connected Layer for Convolutional Neural Network

D Yang, J Cao, YZ Ma, J Yu, S Jiang, L Zhou - International Conference on …, 2023 - Springer
The fully connected (FC) layer is generally located behind the global pooling layer in the
convolutional neural network (CNN). Its essence is the weighted summation of the features …

TDRConv: Exploring the Trade-off Between Feature Diversity and Redundancy for a Compact CNN Module

H Hu, D Zhou, H Xu, Q Chen, Q Guan… - … Conference on Intelligent …, 2023 - Springer
Rich or even redundant features without losing diversity of feature maps can undoubtedly
help to improve network performance. In this work, we propose a compact CNN module by …