LiteDepthwiseNet: An Extreme Lightweight Network for Hyperspectral Image Classification

TL;DR

LiteDepthwiseNet is a highly efficient hyperspectral image classification network that uses 3D depthwise convolution, removes certain layers to reduce overfitting, and employs focal loss to improve performance on small datasets.

Contribution

The paper introduces LiteDepthwiseNet, a novel lightweight architecture for hyperspectral image classification that reduces parameters and computational cost while enhancing accuracy on small datasets.

Findings

Achieves state-of-the-art performance on benchmark datasets.

Uses minimal parameters and low computational cost.

Improves overfitting issues by removing ReLU and Batch Normalization layers.

Abstract

Deep learning methods have shown considerable potential for hyperspectral image (HSI) classification, which can achieve high accuracy compared with traditional methods. However, they often need a large number of training samples and have a lot of parameters and high computational overhead. To solve these problems, this paper proposes a new network architecture, LiteDepthwiseNet, for HSI classification. Based on 3D depthwise convolution, LiteDepthwiseNet can decompose standard convolution into depthwise convolution and pointwise convolution, which can achieve high classification performance with minimal parameters. Moreover, we remove the ReLU layer and Batch Normalization layer in the original 3D depthwise convolution, which significantly improves the overfitting phenomenon of the model on small sized datasets. In addition, focal loss is used as the loss function to improve the model's attention on difficult samples and unbalanced data, and its training performance is significantly better than that of cross-entropy loss or balanced cross-entropy loss. Experiment results on three benchmark hyperspectral datasets show that LiteDepthwiseNet achieves state-of-the-art performance with a very small number of parameters and low computational cost.