电子学报2025，Vol.53Issue(3)：962-973,12.DOI:10.12263/DZXB.20240754
基于坐标重要性池化和解耦类别对齐蒸馏的图像分类算法
Image Classification Algorithm Based on Coordinate Importance Pooling and Decoupled Class Alignment Distillation
摘要
Abstract
An image classification algorithm based on coordinate importance pooling and decoupled class alignment distillation is proposed to improve the image classification accuracy of convolutional neural networks while achieving net-work lightweighting.Firstly,a coordinate importance pooling module is designed and embedded it into ResNet34,in order to fully utilize the positional information of image pixels to enhance the ability to discriminate important features.Secondly,BlurPool is used to mitigate the impact on network performance due to shift equivariance during down-sampling,and to construct the teacher network.Finally,the decoupled class alignment distillation algorithm was constructed to efficiently mi-grate image classification knowledge from the teacher network to the lightweight MobileNetV3 network,which considers the knowledge of target and non-target class separately and introduces correlation information between the class.The experi-mental results on different datasets showed that the proposed teacher network effectively improves the classification perfor-mance,and the distillation-trained student network achieves superior overall performance than other networks of the same magnitude,making it better applicable to practical scenarios with limited computational and storage power.关键词
图像分类/轻量化/知识蒸馏/ResNet34/坐标重要性池化Key words
image classification/lightweight/knowledge distillation/ResNet34/coordinate importance pooling分类
信息技术与安全科学引用本文复制引用
刘颖,薛家昊,张伟东,许志杰..基于坐标重要性池化和解耦类别对齐蒸馏的图像分类算法[J].电子学报,2025,53(3):962-973,12.基金项目
国家自然科学基金(No.62106195) National Natural Science Foundation of China(No.62106195) （No.62106195）
