RFAssigner: A Generic Label Assignment Strategy for Dense Object Detection
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866914232927256576 |
|---|---|
| author | Guan, Ziqian Fu, Xieyi Wang, Yuting Xiao, Haowen Zhu, Jiarui Zhu, Yingying Liu, Yongtao Gu, Lin |
| author_facet | Guan, Ziqian Fu, Xieyi Wang, Yuting Xiao, Haowen Zhu, Jiarui Zhu, Yingying Liu, Yongtao Gu, Lin |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_01240 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | RFAssigner: A Generic Label Assignment Strategy for Dense Object Detection Guan, Ziqian Fu, Xieyi Wang, Yuting Xiao, Haowen Zhu, Jiarui Zhu, Yingying Liu, Yongtao Gu, Lin Computer Vision and Pattern Recognition 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. |
| title | RFAssigner: A Generic Label Assignment Strategy for Dense Object Detection |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2601.01240 |