Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Awesome-RAG-Vision: a curated list of advanced retrieval augmented generation (RAG) for Computer Vision
| Date | Stars |
|---|---|
| 2026-08-04 | 339 |
| 2026-08-18 | 341 |
| 2026-08-19 | 342 |
| 2026-08-28 | 343 |
| 2026-09-08 | 342 |
| 2026-09-15 | 342 |
| 2026-09-17 | 341 |
| 2026-09-24 | 341 |
| 2026-09-30 | 340 |
| 2026-10-07 | 341 |
| 2026-10-15 | 341 |
| 2026-10-17 | 340 |
| 2026-10-22 | 340 |
| 2026-11-04 | 340 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
<!-- Generated by scripts/render_catalog.py from data/papers.json. --> <div align="center"> # Awesome RAG in Computer Vision **Retrieval-augmented understanding, generation and embodied intelligence**    [Survey](https://arxiv.org/abs/2503.18016) · [Paper list](#paper-list) · [JSON](data/papers.json) · [CSV](data/papers.csv) · [BibTeX](data/papers.bib) · [Contribute](CONTRIBUTING.md) </div> A curated research guide to **how retrieved evidence, exemplars, visual references and memory support vision models**. Covers images, documents, video, medical applications, visual synthesis, 3D and embodied agents. **175 unique papers · 114 proceedings/journal papers · 61 preprints · 12 categories.** Every entry has a direct DBLP record. A linked proceedings page supplies formal venue evidence where only the preprint is indexed in DBLP. Conference and preprint versions count once. <p align="center"><img src="assets/acl-figures/overview.png" alt="Four roles of retrieved information: evidence, exemplars, synthesis references and embodied memory" width="100%"></p> *From the survey revision: four roles of retrieved information. [More figures and editable sources](FIGURES.md).* ## Start reading These are entry points for different questions, rather than a performance ranking. | Your question | Suggested entry points | What to look for | |---|---|---| | How does multimodal RAG differ from text-only RAG? | [MuRAG](https://aclanthology.org/2022.emnlp-main.375) · [UniversalRAG](https://aclanthology.org/2026.acl-long.177/) | Modalities and granularity of the retrieved evidence | | How do we retrieve document pages? | [VisRAG](https://openreview.net/forum?id=zG459X3Xge) · [ColPali](https://openreview.net/forum?id=ogjBpZ8uSi) | Separate the downstream answer pipeline from its retrieval component | | How is video evidence selected? | [VideoRAG](https://aclanthology.org/2025.findings-acl.1096/) · [Video-RAG](http://papers.nips.cc/paper_files/paper/2025/hash/f593c9c251d4d7cf14d4ab9861dfb7eb-Abstract-Conference.html) | Retrieval over a video corpus versus a supplied video | | How do we evaluate visual retrieval augmentation? | [MRAG-Bench](https://openreview.net/forum?id=Usklli4gMc) · [REAL-MM-RAG](https://aclanthology.org/2025.acl-long.1528/) | Candidate collections, relevance labels and answer evaluation | | How can retrieval guide visual synthesis? | [Re-Imagen](https://openreview.net/forum?id=XSEBx0iSjFQ) · [ImageRAG](https://proceedings.iclr.cc/paper_files/paper/2026/hash/136729ae4b0fee25a0d28077442506da-Abstract-Conference.html) | Retrieved visual references and generation conditioning | | How does agent memory support actions? | [Embodied-RAG](https://doi.org/10.48550/ARXIV.2409.18313) · [TrajRAG](https://openaccess.thecvf.com/content/CVPR2026/html/Wang_TrajRAG_Retrieving_Geometric-Semantic_Experience_for_Zero-Shot_Object_Navigation_CVPR_2026_paper.html) | Memory construction, retrieval and the downstream action | | What changes in medical applications? | [MMed-RAG](https://openreview.net/forum?id=s5epFPdIW6) · [RULE](https://aclanthology.org/2024.emnlp-main.62) | Clinical evidence, reliability and domain-specific retrieval | ## Recent verified additions and publication updates Targeted update: **19 October 2026**. The full list below is sorted by publication/event year, then title. This selection highlights additions and metadata changes; it is not an exhaustive search of all recent work. | Paper | Verified publication | Why it is useful here | |---|---|---| | [UniversalRAG: Retrieval-Augmented Generation over Corpora of Diverse Modalities and Granularities](https://aclanthology.org/2026.acl-long.177/) | ACL 2026 · [DBLP](https://dblp.org/rec/conf/acl/YeoKJBH26) | Unified retrieval across modalities and granularities; merged with its d
Excerpt of 64,594 characters
Read on GitHubWould you bet a product on this? Bounded 0–100 and slow moving.
matched fp:bbead4640b4522d7, desc:retrieval-augmented generation, desc:retrieval augmented