AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

Collaborative Training between Region Proposal Localization and Classification for Domain Adaptive Object Detection

Ganlong Zhao, Guanbin Li, Ruijia Xu, Liang Lin

ECCV 2020 · Detection and segmentation › Open-set and few-shot detection · approach: Domain-adaptive object detection · 126 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 collaborative self-training between RPN and RPC (ablation)source-only Faster RCNN · Sim10k to Cityscapes (car)AP on car: 34.57 -> 42.33Improved
add naive local alignment (ablation)source-only Faster RCNN · Sim10k to Cityscapes (car)AP on car: 34.57 -> 37.03Improved
use RPN weighted local alignment instead of naive local alignment (ablation)source-only Faster RCNN · Sim10k to Cityscapes (car)AP on car: 37.03 -> 38.31Improved
add MCD alone on RPN/RPC discrepancy (ablation)source-only Faster RCNN · Sim10k to Cityscapes (car)AP on car: 34.57 -> 36.42No clear effect
add weighted local alignment to CST (ablation)CST model · Sim10k to Cityscapes (car)AP on car: 42.33 -> 43.08Improved
add MCD to weighted alignment plus CST (ablation)weighted alignment + CST · Sim10k to Cityscapes (car)AP on car: 43.08 -> 44.51Improved
add all three components (CST, weighted alignment, MCD) (ablation)source-only Faster RCNN · Sim10k to Cityscapes (car)AP on car: 34.57 -> 44.51Improved
use proposed method instead of best prior (SCDA)SCDA · Cityscapes to FoggyCityscapesmAP: 33.8 -> 35.9Improved
use proposed method instead of best prior (SCDA)SCDA · Sim10k to Cityscapes (car)AP on car: 43.05 -> 44.51Improved
use proposed method instead of best prior (SCDA)SCDA · KITTI to Cityscapes (car)AP on car: 42.5 -> 43.6Improved

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