arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2601.01240cs.CV
收藏

RFAssigner: 一种用于密集目标检测的通用标签分配策略

RFAssigner: A Generic Label Assignment Strategy for Dense Object Detection

Guangzhou Institutes of Biomedicine and Health, Chinese Academy of Sciences(广州生物医学研究院,中国科学院);North China Institute of Science and Technology(华北科学技术学院);RIKEN AIP, Japan(日本RIKEN AIP)

Ziqian Guan, Xieyi Fu, Yuting Wang, Haowen Xiao, Jiarui Zhu, Yingying Zhu, Yongtao Liu, Lin Gu

AI总结:

RFAssigner通过多尺度学习增强策略,提升密集目标检测器在不同物体尺度上的性能表现。

摘要:

标签分配是训练密集目标检测器的关键组成部分。当前最先进的方法通常为每个训练样本分配正样本和负样本权重,并在训练过程中优化分配方案。然而,这些策略往往对小对象分配的正样本数量不足,导致训练过程中的尺度失衡。为了解决这一限制,我们引入了RFAssigner,一种新的分配策略,旨在增强密集检测器的多尺度学习能力。RFAssigner首先使用点基先验建立初始正样本集。然后利用高斯感受场(GRF)距离来衡量未分配候选位置与真实对象的GRF之间的相似性。基于这一度量,RFAssigner从未分配池中自适应地选择补充正样本,促进不同物体尺度之间的更平衡学习过程。在三个具有不同物体尺度分布的数据集上的全面实验验证了我们方法的有效性和通用性。值得注意的是,一个配备RFAssigner的单FCOS-ResNet-50检测器在所有物体尺度上均实现了最先进的性能,持续优于现有策略,无需辅助模块或启发式方法。

英文摘要:

Label assignment is a critical component in training dense object detectors. State-of-the-art methods typically assign each training sample a positive and a negative weight, optimizing the assignment scheme during training. However, these strategies often assign an insufficient number of positive samples to small objects, leading to a scale imbalance during training. To address this limitation, we introduce RFAssigner, a novel assignment strategy designed to enhance the multi-scale learning capabilities of dense detectors. RFAssigner first establishes an initial set of positive samples using a point-based prior. It then leverages a Gaussian Receptive Field (GRF) distance to measure the similarity between the GRFs of unassigned candidate locations and the ground-truth objects. Based on this metric, RFAssigner adaptively selects supplementary positive samples from the unassigned pool, promoting a more balanced learning process across object scales. Comprehensive experiments on three datasets with distinct object scale distributions validate the effectiveness and generalizability of our method. Notably, a single FCOS-ResNet-50 detector equipped with RFAssigner achieves state-of-the-art performance across all object scales, consistently outperforming existing strategies without requiring auxiliary modules or heuristics.

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。

↑