Memory-based Error Label Suppression for Embodied Self-Improving Object Detection

CASIA OpenIR > 中科院工业视觉智能装备工程实验室 > 先进制造与自动化

	Memory-based Error Label Suppression for Embodied Self-Improving Object Detection
	Deng JR(邓杰仁)1,2 ; Zhang HJ(张好剑)1 ; Hu JH(胡建华)1 ; Wang YK(王云宽)1
	2024-06-04
会议名称	2024 IEEE 20th International Conference on Automation Science and Engineering
会议日期	2024-8-28
会议地点	意大利巴里
出版者	IEEE
摘要	This paper introduces a novel method for embodied self-improving object detection, aimed at enhancing the object detection model by gathering additional labeled samples post-pre-training without the need for human supervision. Current self-improving strategies autonomously label new samples using 3D consistency, yet they often incorporate a substantial amount of mislabeled samples, thereby diminishing the potential performance improvement to the model. To counter this issue, we propose a memory-based method for suppressing error labels to minimize their adverse impact. This error label suppression mechanism includes LoRA output constraint and exemplar prototype constraints, which leverage explicit memories of correct prototypes and implicit memories of correctly learned parameters, respectively. These mechanisms effectively reduce the negative effects of erroneous labels on the model’s learning process. Building upon the robustness provided by our Memory-Based Error Label Suppression, we further incorporates Cross-View Redundant Labeling to introduce a higher quantity of accurate samples, thus amplifying the benefits of embodied self-improving. Experimental results demonstrate that our method exhibits more robustness against erroneous samples compared to existing methods, leading to significantly performance improvement.
收录类别	EI
语种	英语
七大方向——子方向分类	智能机器人
国重实验室规划方向分类	实体人工智能系统感认知
是否有论文关联数据集需要存交	否
文献类型	会议论文
条目标识符	http://ir.ia.ac.cn/handle/173211/57431
专题	中科院工业视觉智能装备工程实验室_先进制造与自动化
通讯作者	Hu JH(胡建华)
作者单位	1.中国科学院自动化研究所 2.中国科学院大学
第一作者单位	中国科学院自动化研究所
通讯作者单位	中国科学院自动化研究所
推荐引用方式 GB/T 7714	Deng JR,Zhang HJ,Hu JH,et al. Memory-based Error Label Suppression for Embodied Self-Improving Object Detection[C]:IEEE,2024.

条目包含的文件		下载所有文件
文件名称/大小	文献类型	版本类型	开放类型	使用许可
root.pdf（2603KB）	会议论文		开放获取	CC BY-NC-SA	浏览下载