Enhanced Knowledge Distillation for Detection Transformer via Teacher Prediction Refinement

The paper introduces Teacher Prediction Refinement Distillation (TPRD), a plug‑in module for Detection Transformers that refines teacher predictions before distillation. TPRD corrects degraded positive predictions and suppresses overconfident negatives, while preserving informative dark knowledge through Maximum Dark Knowledge Preservation. Experiments on MS COCO and PASCAL VOC show that these refinements improve the quality of supervision and the resulting student model’s performance.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Computer Vision.

• llms
• computer-vision
• efficiency