论文类型：期刊论文
论文编号：s11063-022-10855-0
第一作者：Xiao Wang
通讯作者：Baochang Zhang
合写作者：Xiang Xiang,Xuhui Liu,Jianying Zheng,QingLei Hu
发表刊物：Neural Processing Letters
收录刊物：SCI、EI
所属单位：Springer Nature
刊物所在地：比利时
学科门类：工学
文献类型：J
关键字：Active Learning, Weak Supervision, Object Detection
DOI码：10.1007/s11063-022-10855-0
发表时间：2022-05-30
影响因子：3.0
摘要：Weakly supervised object detection which reduces the need for strong supervision during training has recently made significant achievements. However, it remains a challenging issue due to the time-consuming and labor-intensive problems in application. To further reduce the label cost, we introduce a new fusion method of weakly supervised learning and active learning in a unified framework for object detection. Weakly supervised learning based on min-entropy latent model is used to weaken the labels by image-label, while active learning is used to reduce the quantity of labelled images. The fusion method proposed can effectively reduce the dependency of object detection on manual annotation. In this paper, we introduce three strategies of active learning, including least confidence sampling, margining sampling and weighted classification sampling. To validate the effectiveness of each strategy and different sample compositions in weakly supervised learning object detection, we conducted lots of experiments. Extensive experiments show that the combination of image-level labeling and active learning can achieve comparable results with the previous state-of-the-art methods with much lower label cost.
发布期刊链接：https://link.springer.com/article/10.1007/s11063-022-10855-0