收缩、分离和聚合:面向长尾视觉识别的特征平衡方法  

Shrink,Separate and Aggregate:A Feature Balancing Method for Long-tailed Visual Recognition

作　　者：杨佳鑫 于淼淼 李虹颖 李硕豪 范灵毓 张军[1,2] YANG Jia-Xin;YU Miao-Miao;LI Hong-Ying;LI Shuo-Hao;FAN Ling-Yu;ZHANG Jun(College of System Engineering,National University of Defense Technology,Changsha 410073;Laboratory for Big Data and Decision,Changsha 410073;Unit 96962 of the PLA,Beijing 102206)

机构地区：[1]国防科技大学系统工程学院,长沙410073 [2]大数据与决策实验室,长沙410073 [3]中国人民解放军96962部队,北京102206

出　　处：《自动化学报》2024年第5期898-910,共13页Acta Automatica Sinica

基　　金：国家自然科学基金(62101571);湖南省自然科学基金(2021JJ40685)资助。

摘　　要：数据在现实世界中通常呈现长尾分布,即,少数类别拥有大量样本,而多数类别仅有少量样本.这种数据不均衡的情况会导致在该数据集上训练的模型对于样本数量较少的尾部类别产生过拟合.面对长尾视觉识别这一任务,提出一种面向长尾视觉识别的特征平衡方法,通过对样本在特征空间中的收缩、分离和聚合操作,增强模型对于难样本的识别能力.该方法主要由特征平衡因子和难样本特征约束两个模块组成.特征平衡因子利用类样本数量来调整模型的输出概率分布,使得不同类别之间的特征距离更加均衡,从而提高模型的分类准确率.难样本特征约束通过对样本特征进行聚类分析,增加不同类别之间的边界距离,使得模型能够找到更合理的决策边界.该方法在多个常用的长尾基准数据集上进行实验验证,结果表明不但提高了模型在长尾数据上的整体分类精度,而且显著提升了尾部类别的识别性能.与基准方法BS相比较,该方法在CIFAR100-LT、ImageNet-LT和iNaturalist 2018数据集上的性能分别提升了7.40%、6.60%和2.89%.Data in the real world often exhibits a long-tailed distribution,where a few classes have a large number of samples,while most classes have only a few samples.This data imbalance can lead to overfitting in the model trained on this dataset for tail classes with fewer samples.To address this problem,we propose a feature balancing method for long-tailed visual recognition,which enhances the model's ability to recognize hard samples by shrinking,separating and aggregating samples in the feature space.The method consists of two modules:Feature balance factor and hard sample feature constraint.The feature balance factor uses the sample number of classes to adjust the model's output probability distribution,making the feature distance between different classes more balanced,thereby improving the model's classification accuracy.The hard sample feature constraint performs clustering analysis on the sample features,increasing the boundary distance between different classes,enabling the model to find a more reasonable decision boundary.We conduct experiments on several common long-tailed benchmark datasets,experimental results show that the proposed method not only improves the model's overall classification accuracy on long-tailed data,but also significantly enhances the recognition performance of tail classes.Compared with baseline method BS,the proposed method achieves performance improvements of 7.40%,6.60% and 2.89% on CIFAR100-LT,ImageNet-LT and iNaturalist 2018 datasets respectively.

关 键 词：长尾识别 损失设计 特征平衡 特征约束 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]