Research map / Detection and segmentation / Open-set and few-shot detection
Task-Agnostic Incremental Vision-Language Object Detection via Prompt Augmentation and Distribution-Aware Fusion
Yonghan Jiang, Zhengyuan Xie, Wenchu Liu, Linlan Huang, Fei Yang, Xialei Liu
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 changed | Where | Measured | Effect |
|---|---|---|---|
| replace static averaging with TTDF distribution-aware fusion (ablation) | DitHub · ODinW-13 TA-IVLOD | Avg mAP: 32.5 -> 45.9 | Improved |
| add SPA on top of TTDF (ablation) | TADA · ODinW-13 TA-IVLOD | Avg mAP: 45.9 -> 64.5 | Improved |
| combine learned classes with noise tokens in prompt augmentation versus noise tokens only (ablation) | SPA · ODinW-13 TA-IVLOD | Avg mAP: 60.3 -> 64.5 | Improved |
| use synthetic noise tokens instead of COCO classes as distractors (ablation) | SPA · ODinW-13 TA-IVLOD | Zcoco mAP: 46.0 -> 47.4 | Improved |
| use Mahalanobis instead of Euclidean distance (ablation) | TTDF · ODinW-13 TA-IVLOD | Avg mAP: 51.3 -> 64.5 | Improved |
| use Mahalanobis instead of Euclidean distance (ablation) | TTDF · ODinW-13 TA-IVLOD | Zcoco mAP: 45.9 -> 47.2 | Improved |
| add TADA to DitHub in standard IVLOD | DitHub · ODinW-13 standard IVLOD | Avg mAP: 66.3 -> 67.5 | Improved |
| add TADA to DitHub in TA-IVLOD | DitHub · ODinW-13 TA-IVLOD | Avg mAP: 32.5 -> 64.5 | Improved |
| add TADA in 10-shot TA-IVLOD | DitHub · ODinW-13 10-shot TA-IVLOD | AvgTA mAP: 30.3 -> 53.9 | Improved |
| add TADA versus GCD under conventional IOD | GCD · MS COCO IOD (40-40) | mAP: 55.2 vs 54.5 | No clear effect |
| switch from single-task to cross-task prompts | DitHub · ODinW-13 | Avg mAP: 66.3 -> 32.5 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DitHub — DitHub: A Modular Framework for Incremental Open-Vocabulary Object Detection (NeurIPS 2025) · better on 7 of 7 shared results
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
- Zero-shot Generalizable Incremental Learning for Vision-Language Object Detection (NeurIPS 2024 · similarity 0.92)
- GCD: Advancing Vision-Language Models for Incremental Object Detection via Global Alignment and Correspondence Distillation (AAAI 2025 · similarity 0.88)
- DitHub: A Modular Framework for Incremental Open-Vocabulary Object Detection (NeurIPS 2025 · similarity 0.87)
- Roboflow100-VL: A Multi-Domain Object Detection Benchmark for Vision-Language Models (NeurIPS 2025 · similarity 0.87)
- Cyclic Contrastive Knowledge Transfer for Open-Vocabulary Object Detection (ICLR 2025 · similarity 0.87)
- Learning Task-Aware Language-Image Representation for Class-Incremental Object Detection (AAAI 2024 · similarity 0.87)
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.
