基于数据的不平衡学习大规模分层的目标检测网络方案

Learning an Efficient Network for Large-Scale Hierarchical Object Detection with Data Imbalance

通过对公开图形数据集的详细分析,发现数据集有四种典型特征:规模大、标签系统分层、标注不完整和数据不平衡。针对以上数据集的典型特征,在采用更大的骨干网、分布式Softmax损失函数、分类别采样策略、专家模型和重分类器等应对策略后,单模型mAP精度最优能达到62.29。经过集成之后,mAP精度能最终可提升到67.17。试验结果表明,基于数据的不平衡学习大规模分层的目标检测网络方案有效地提升目标的识别率和准确率。

This report details based on our detailed analysis on the Open Images dataset,it is found that there are four typical features:large-scale,hier-archical tag system,severe annotation incompleteness and data imbalance.Considering these characteristics,many strategies are employed,including larger backbone,dis-tributed softmax loss,class-aware sampling,expert model,and heavier classifier.In virtue of these effective strategies,our best single model could achieve a mAP of 61.90.After ensemble,the final mAP is boosted to 67.17 and 64.21.