arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2302.11299cs.CV
收藏

面向单阶段目标检测的端到端半监督学习

Towards End-to-end Semi-supervised Learning for One-stage Object Detection

Xiamen University(厦门大学)

Gen Luo, Yiyi Zhou, Lei Jin, Xiaoshuai Sun, Rongrong Ji

摘要:

半监督目标检测(SSOD)是计算机视觉的一个研究热点,它可以大大减少对昂贵的边界框标注的需求。尽管取得了巨大的成功,但现有的进展主要集中在FasterRCNN等两阶段检测网络上,而对一阶段检测器的研究往往被忽视。在本文中,我们专注于半监督学习的先进和流行的一级检测网络YOLOv 5。与Faster-RCNN相比,YOLOv 5的实现要复杂得多,YOLOv 5中使用的各种训练技术也会降低SSOD的收益。除了这个挑战,我们还揭示了两个关键问题,在一阶段SSOD,这是低质量的伪标记和多任务优化冲突,分别。为了解决这些问题,我们提出了一种新的师生学习配方称为OneTeacher,具有两个创新的设计,即多视图伪标签细化(MPR)和解耦半监督优化(DSO)。特别是,MPR通过增强视图细化和全局视图过滤来提高伪标签的质量,DSO通过结构调整和特定于任务的伪标签来处理联合优化冲突。此外,我们还对YOLOv 5的实施进行了认真修改,以最大限度地发挥SSOD的效益,这也与现有的SSOD方法进行了公平比较。为了验证OneTeacher,我们在COCO和Pascal VOC上进行了广泛的实验。大量的实验表明,OneTeacher不仅可以实现优于比较方法的性能,例如,与无偏见教师相比,AP相对增益为15.0%,但也很好地处理了一阶段SSOD中的关键问题。我们的源代码可以在https://github.com/luogen1996/OneTeacher上找到。

英文摘要:

Semi-supervised object detection (SSOD) is a research hot spot in computer vision, which can greatly reduce the requirement for expensive bounding-box annotations. Despite great success, existing progress mainly focuses on two-stage detection networks like FasterRCNN, while the research on one-stage detectors is often ignored. In this paper, we focus on the semi-supervised learning for the advanced and popular one-stage detection network YOLOv5. Compared with Faster-RCNN, the implementation of YOLOv5 is much more complex, and the various training techniques used in YOLOv5 can also reduce the benefit of SSOD. In addition to this challenge, we also reveal two key issues in one-stage SSOD, which are low-quality pseudo-labeling and multi-task optimization conflict, respectively. To address these issues, we propose a novel teacher-student learning recipe called OneTeacher with two innovative designs, namely Multi-view Pseudo-label Refinement (MPR) and Decoupled Semi-supervised Optimization (DSO). In particular, MPR improves the quality of pseudo-labels via augmented-view refinement and global-view filtering, and DSO handles the joint optimization conflicts via structure tweaks and task-specific pseudo-labeling. In addition, we also carefully revise the implementation of YOLOv5 to maximize the benefits of SSOD, which is also shared with the existing SSOD methods for fair comparison. To validate OneTeacher, we conduct extensive experiments on COCO and Pascal VOC. The extensive experiments show that OneTeacher can not only achieve superior performance than the compared methods, e.g., 15.0% relative AP gains over Unbiased Teacher, but also well handle the key issues in one-stage SSOD. Our source code is available at: https://github.com/luogen1996/OneTeacher.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑