AN IQE CRITERION-BASED METHOD FOR SAR IMAGES CLASSIFICATION NETWORK PRUNING

被引：1

作者：

Wang, Jielei ^{[1
]}

Cui, Zongyong ^{[1
]}

Cao, Zongjie ^{[1
]}

Wang, Hanzeng ^{[1
]}

Cao, Changjie ^{[1
]}

机构：

[1] Univ Elect Sci & Technol China, Sch Informat & Commun Engn, Chengdu, Peoples R China

来源：

2021 IEEE INTERNATIONAL GEOSCIENCE AND REMOTE SENSING SYMPOSIUM IGARSS | 2021年

基金：

中国国家自然科学基金;

关键词：

Deep convolutional neural network; SAR-ATR; Network pruning; DCNN acceleration; Knowledge Processing Unit (KPU);

D O I：

10.1109/IGARSS47720.2021.9553292

中图分类号：

P [天文学、地球科学];

学科分类号：

07 ;

摘要：

Deep convolutional neural networks (DCNNs) have been widely used for SAR image target recognition. However, the huge demands of DCNNs for computing, storage, and energy resources limit their use on edge computing devices. In this article, we propose a method based on image quality evaluation (IQE) criterion to prune deep neural networks. We use IQE criterion to identify unimportant filters, and then remove them, to obtain a lightweight network while maintaining the performance of the neural network as much as possible. Besides, we verified the effectiveness of our method on the MSTAR dataset with cheap edge computing devices.

引用

页码：3593 / 3596

页数：4

共 10 条

[1]

Bolukbasi Tolga, P 34 INT C MACH LEAR, V70, P527

[2]

Han S., Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding

[3]

Hu H., 2016, ARXIV160703250

[4]

Li H., Pruning Filters for Efficient ConvNets

[5]

Liu Congcong, CHANNEL PRUNING BASE, V156, P84

[6]

Luo Jian-Hao, ENTROPY BASED PRUNIN

[7]

Mao Huizi, EXPLORING REGULARITY

[8]

Molchanov Pavlo., 2016, Pruning convolutional neural networks for resource efficient transfer learning

[9]

Simonyan K, 2015, Arxiv, DOI arXiv:1409.1556

[10]

Wen W., Advances in Neural Information Processing Systems, P2074

← 1 →