Showing cs.CVShow all
2 papers · 1 filter
cs.CV2025
Aligned Contrastive Loss for Long-Tailed Recognition
Jiali Ma, Jiequan Cui, Maeno Kazuki +4
In this paper, we propose an Aligned Contrastive Learning (ACL) algorithm to address the long-tailed recognition problem. Our findings indicate that while multi-view training boost…
cs.CV2024
Unbiased Regression Loss for DETRs
Edric, Ueta Daisuke, Kurokawa Yukimasa +2
In this paper, we introduce a novel unbiased regression loss for DETR-based detectors. The conventional regression loss tends to bias towards larger boxes, as they dispropo…