Mixed-Precision Network Quantization for Infrared Small Target Segmentation
- 1 January 2024
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 62 (01962892) , 1-12
Abstract
Network quantization is leveraged to reduce the model size, memory footprint, and computational cost of deep neural networks. It is achieved by representing float weights and activations with lower bit counterparts, which is essential for model deployment on resource-limited devices. However, due to the extremely small size of infrared small targets in the feature map, low-bit quantization could lead to huge information loss of small targets and thus causes severe segmentation performance degradation. To achieve low-bit quantization while maintaining the segmentation performance, we first study the quantization sensitivity of small target segmentation network and observe the sensitivity heterogeneity of different layers in the network. Specifically, feature maps in shallow layers and encoder subnetwork are more vulnerable to information loss caused by quantization as compared to deep layers and decoder subnetwork. Based on these observations, we are motivated to assign a different bitwidth for each block according to their quantization sensitivity. A simple yet effective symmetrically progressive decreasing mixed-precision quantization (SPMix-Q) method is proposed to achieve high-performance segmentation under low-bit quantization (i.e., 2.42 bits for weights and 3.82 bits for activations). The experimental results show that our SPMix-Q achieves comparable accuracy with only 1/13 model size, 1/4.6 memory footprint, and 1/29 computational cost to the full-precision counterparts. Compared with the homogeneous low-bit quantization methods, our method achieves much better performance in terms of intersection of union (IoU) on the benchmark datasets. Our mobile-system-on-a-chip (SOC) (e.g., Kyrin 980, Snapdragon 660, and Dimensity 800U) deployable android application package (APK) is available at: https://github.com/YeRen123455/SIRST-Quantization-Deployment .Keywords
Funding Information
- National Key Research and Development Program of China (2021YFB3100800)
- National Natural Science Foundation of China (62001478)
This publication has 46 references indexed in Scilit:
- Combining Deep Denoiser and Low-rank Priors for Infrared Small Target DetectionPattern Recognition, 2022
- SVNet: Where SO(3) Equivariance Meets Binarization on Point Cloud Representation2022
- Mixed-precision quantized neural networks with progressively decreasing bitwidthPattern Recognition, 2021
- U-Net Fixed-Point Quantization for Medical Image Segmentation2019
- Fully Quantized Network for Object Detection2019
- Learning to Quantize Deep Networks by Optimizing Quantization Intervals With Task Loss2019
- XNOR-Net: ImageNet Classification Using Binary Convolutional Neural Networks2016
- U-Net: Convolutional Networks for Biomedical Image Segmentation2015
- Infrared dim target detection based on visual attentionInfrared Physics & Technology, 2012
- Scale invariant small target detection by optimizing signal-to-clutter ratio in heterogeneous background for infrared search and trackPattern Recognition, 2012