AI Paper Atlas

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

Mind the Gap: Transferring Labels to Align Object Detection Datasets

Mikhail Kennerley, Angelica I. Aviles-Rivero, Carola-Bibiane Schönlieb, Robby T. Tan

CVPR 2026 · Detection and segmentation › Weakly supervised detection and segmentation · approach: Pseudo-labeling for semi-supervised detection

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 class-aware SFF attention instead of no attention (ablation)LAT (FRCNN) · CityscapesAP: 57.5 -> 60.1Improved
use class-aware SFF instead of standard attention (ablation)LAT (FRCNN) · WaymoAP: 47.0 -> 48.5Improved
apply scaling to class branch rather than feature branch (ablation)LAT (FRCNN) · CityscapesAP: 59.6 -> 60.1Improved
add attention scaling versus none (ablation)LAT (FRCNN) · CityscapesAP: 57.8 -> 60.1Improved
fine-tune on source-only batches after mixed pretraining (ablation)LAT downstream detector · CityscapesAP: 58.9 -> 60.1Improved
50/50 batches instead of mixed batches (ablation)LAT downstream detector · CityscapesAP: 58.9 -> 57.5Hurt
add SAM3 predictions to LATLAT (RT-DETR) · CityscapesAP: 60.6 -> 61.0Improved
use SAM3 labels alone instead of LATlabel transfer · nuImagesAP: 39.5 -> 32.6Hurt
semi-supervised student-teacher using other datasetsFRCNN baseline · CityscapesAP: 55.2 -> 55.1No clear effect
LAT on large target datasets with standard trainingFRCNN · BDD100KAP: 57.2 -> 56.1Hurt
longer training for LATFRCNN · BDD100KAP: 56.1 -> 57.8Improved
longer training for LATRT-DETR · CityscapesAP: 61.2 -> 59.8Hurt

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