AI Paper Atlas

Research map / Label-efficient and robust learning / Knowledge distillation

G-DetKD: Towards General Distillation Framework for Object Detectors via Contrastive and Semantic-Guided Feature Imitation

Lewei Yao, Renjie Pi, Hang Xu, Wei Zhang, Zhenguo Li, Tong Zhang

ICCV 2021 · Label-efficient and robust learning › Knowledge distillation · approach: Knowledge distillation for detectors · 40 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use SGFI instead of RoI feature imitation (ablation)FasterRCNN-R18 student · COCO valAP: 36.4 -> 36.7Improved
use SGFI instead of GT-mask imitation (ablation)RetinaNet-R18 student · COCO valAP: 34.3 -> 36.0Improved
add contrastive KD on top of SGFI (ablation)FasterRCNN-R18 student · COCO valAP: 36.7 -> 37.6Improved
add prediction KD and head transfer to SGFI+CKD (ablation)FasterRCNN-R18 student · COCO valAP: 37.6 -> 37.9Improved
add prediction KD to SGFI (ablation)RetinaNet-R18 student, FasterRCNN-R50 teacher · COCO valAP: 36.1 -> 36.1No clear effect
use class-aware regression KD instead of class-agnostic (ablation)FasterRCNN-R18 student · COCO valAP: 34.7 -> 35.7Improved
increase contrastive memory queue from 8x1024 to 80x1024 (ablation)CKD FasterRCNN-R18 · COCO valAP: 36.8 -> 37.1Improved
raise IoU threshold for negative assignment from 0.1 to 0.5 (ablation)CKD FasterRCNN-R18 · COCO valAP: 36.8 -> 37.1Improved
use linear instead of nonlinear projection head (ablation)CKD FasterRCNN-R18 · COCO valAP: 36.7 -> 37.1Improved
form contrastive pairs from classification branch instead of localization branch (ablation)CKD RetinaNet-R18 · COCO valAP: 34.3 -> 35.8Improved
distill from a weaker R18 teacher into R50 studentG-DetKD R50 student · COCO valAP: 37.4 -> 37.7Improved
apply G-DetKD to R101 student with R50 teacherG-DetKD · COCO valAP: 39.4 -> 40.4Improved

Later papers that compared with this method

1 later accepted paper reports G-DetKD as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy