AI Paper Atlas

Research map / Detection and segmentation / Weakly supervised detection and segmentation

Spatial Self-Distillation for Object Detection with Inaccurate Bounding Boxes

Di Wu, Pengfei Chen, Xuehui Yu, Guorong Li, Zhenjun Han, Jianbin Jiao

ICCV 2023 · Detection and segmentation › Weakly supervised detection and segmentation · approach: Weakly supervised detection · 26 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
add Spatial Position Self-Distillation (SPSD) bag construction (ablation)SSD-Det · MS-COCO val, 40% noiseAP: 24.6 -> 27.2Improved
add Spatial Identity Self-Distillation (SISD) for proposal selection (ablation)SSD-Det · MS-COCO val, 40% and 20% noiseAP: 27.2 -> 27.6 (40%), 33.0 -> 33.6 (20%)Improved
re-train Faster R-CNN on SSD-Det pseudo labels after end-to-end training (ablation)SSD-Det · MS-COCO val, 40% noiseAP: 27.6 -> 29.3Improved
end-to-end joint training before re-train instead of training only the refiner (ablation)SSD-Det + Re-Train · MS-COCO val, 40% noiseAP: 29.0 -> 29.3Improved
weight SPSD proposals with stage I score (ablation)SPSD · MS-COCO val, 40% noiseAP: 26.0 -> 26.3No clear effect
use SPSD in both stages instead of only stage II (ablation)SPSD · MS-COCO val, 40% noiseAP: 26.3 -> 27.2Improved
increase SPSD modules from two to three (ablation)SSD-Det · MS-COCO val, 40% noiseAP: 27.2 -> 27.0Hurt
subtract or concatenate object and proposal features in SISD ORE (ablation)SISD · MS-COCO val, 40% noiseAP: 27.2 (subtract), 27.2 (concat) versus 27.6 (add)No clear effect
share the two fc layers in SISD (ablation)SISD · MS-COCO val, 40% noiseAP: 27.6 -> 23.0Hurt
re-train advanced detectors on SSD-Det refined boxesDeformable DETR · MS-COCO, 40% noiseAP: 5.0 -> 35.2Improved
use spatial self-distillation instead of MIL refinementOA-MIL · MS-COCO val, 40% noisedrift rate (%): 15.1 -> 1.5Improved
SSD-Det at low noise on VOC compared with OA-MILOA-MIL · VOC 2007, 10% noiseAP50: 77.4 -> 77.1No clear effect

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