.png){kind=icon}
.png){kind=icon}
Return
DeFRCN-PF: Mitigating Knowledge Forgetting in Few-Shot Object Detection via Class Proxy
.png){kind=icon}
OA
AI
F
C
W
DOI：10.1016/j.patrec.2026.04.032.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• Propose DeFRCN-PF to mitigate catastrophic forgetting in G-FSOD tasks. • Introduce Class Proxy Block to consolidate base-class knowledge. • Design dynamic weight fusion to balance base and novel class learning. • Achieve state-of-the-art results on PASCAL VOC and MS COCO benchmarks. • Improve both base and novel AP, enhancing generalization and retention.
Keywords:
DeFRCN-PF
catastrophic forgetting
few-shot object detection
class proxy block
dynamic weight fusion
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
3.3
Papers:
7.8K
Citations:
1.6W
Organization
Cited Papers