课程学习指导下的半监督目标检测框架  

Semi-supervised object detection framework guided by curriculum learning

作　　者：张英俊[1] 李牛牛 谢斌红[1] 张睿 陆望东 ZHANG Yingjun;LI Niuniu;XIE Binhong;ZHANG Rui;LU Wangdong(College of Computer Science and Technology,Taiyuan University of Science and Technology,Taiyuan Shanxi 030024,China;Shanxi Tianhe Cloud Computing Company Limited,Lvliang Shanxi 033000,China)

机构地区：[1]太原科技大学计算机科学与技术学院,太原030024 [2]山西天河云计算有限公司,山西吕梁033000

出　　处：《计算机应用》2024年第8期2326-2333,共8页journal of Computer Applications

基　　金：山西省基础研究计划项目(20210302123216);吕梁市引进高层次科技人才重点研发项目(2022RC08)。

摘　　要：为了提高伪标签的质量,解决半监督目标检测(SSOD)中的确认偏差问题,并针对现有算法中忽视无标注数据复杂性导致错误伪标签的难点,提出一种课程学习(CL)指导下的SSOD框架,该框架主要由ICSD(IoUConfidence-Standard-Deviation)难度测量器和BP(Batch-Package)训练调度器这2个模块组成。其中,ICSD难度测量器综合考虑了伪边界框之间的交并比(IoU)、置信度、类别标签等信息,并引入C_IOU(Checkpoint_IOU)方法评估无标注数据的可靠性;BP训练调度器设计2种高效调度策略,分别从Batch和Package角度出发,优先选择可靠性指标高的无标记数据,实现以CL的方式充分利用整个无标记数据集。在Pascal VOC和MS-COCO数据集上的广泛对比实验结果表明,所提框架不仅适用于现有的SSOD算法,而且检测精度和稳定性都得到显著提升。In order to enhance the quality of pseudo labels,address the issue of confirmation bias in Semi-Supervised Object Detection(SSOD),and tackle the challenge of ignoring complexities in unlabeled data leading to erroneous pseudo labels in existing algorithms,an SSOD framework guided by Curriculum Learning(CL)was proposed.The framework consisted of two modules:the ICSD(IoU-Confidence-Standard-Deviation)difficulty measurer and the BP(Batch-Package)training scheduler.The ICSD difficulty measurer comprehensively considered information such as IoU(Intersection over Union)between pseudo-bounding boxes,confidence,class label,etc.,and the C_IOU(Checkpoint_IOU)method was introduced to evaluate the reliability of unlabeled data.The BP training scheduler designed two efficient scheduling strategies,starting from the perspectives of Batch and Package respectively,giving priority to unlabeled data with high reliability indicators to achieve full utilization of the entire unlabeled data set in the form of course learning.Extensive comparative experimental results on the Pascal VOC and MS-COCO datasets demonstrate that the proposed framework applies to existing SSOD algorithms and exhibits significant improvements in detection accuracy and stability.

关 键 词：半监督学习 目标检测 课程学习 训练策略 难度测量器 训练调度器 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]