AI Paper Atlas

Research map / Continual learning / Class-incremental

Orthogonal Knowledge Refreshing for Domain-Incremental Object Detection

Aoting Zhang, Dongbao Yang, Chang Liu, Xiaopeng Hong, Can Ma, Yu ZHOU

ECCV 2026 · Continual learning › Class-incremental · approach: Incremental object detection with replay · 2 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 backbone fine-tuning (ablation)Baseline detector · Pascal VOC series Session 4mAP@0.5: 44.2 -> 49.5Improved
Add low-rank subspace expansion (SE) (ablation)FT-Back · Pascal VOC series Session 4mAP@0.5: 49.5 -> 60.9Improved
Add gradient-based orthogonal refreshing (GOR) (ablation)FT-Back + SE · Pascal VOC series Session 4mAP@0.5: 60.9 -> 64.9Improved
Add topology-aware consistency (TAC) (ablation)FT-Back + SE · Pascal VOC series Session 4mAP@0.5: 60.9 -> 63.3Improved
Add TAC without SE (ablation)FT-Back · Pascal VOC series Session 4mAP@0.5: 49.5 -> 60.2Improved
Add TAC on top of SE and GOR (ablation)FT-Back + SE + GOR · Pascal VOC series Session 4mAP@0.5: 64.9 -> 65.2Improved
Add GOR on top of SE and TAC (ablation)FT-Back + SE + TAC · Pascal VOC series Session 4mAP@0.5: 63.3 -> 65.2Improved
Use smaller LoRA ranks (8 instead of 16) (ablation)OKR · Pascal VOC seriesmAP@0.5Hurt
Increase LoRA rank beyond 16 (to 64) (ablation)OKR · Pascal VOC seriesmAP@0.5No clear effect
Add orthogonal projection (GOR) to training (ablation)OKR · Training loss convergencetraining lossNo clear effect
Use OKR instead of LDB+SOYOExemplar-free DIOD · Pascal VOC series Session 4mAP@0.5: 59.6 -> 65.2Improved
Use OKR instead of LDB+SOYOExemplar-free DIOD · BDD100K series Session 3mAP@0.5: 51.7 -> 58.2Improved

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