KwanWaiPang/Awesome-Transformer-based-SLAM

What it solves

Researchers and practitioners need a quick way to find and compare the growing body of work that applies Transformer models to simultaneous localization and mapping (SLAM). Instead of searching scattered sources, this list gathers relevant papers in one place, organized by task so you can see what has been done and where the field is heading.

How it works

The repository is a structured, hand-maintained table of papers. Each entry includes the year, venue, paper title, a link to the arXiv page, any available code repository, and a short note summarizing the approach. The papers are grouped into categories such as full SLAM systems, pose tracking, optical flow, view matching, and mapping. The list is open to contributions via pull requests or issue reports, making it a community-maintained resource.

Who it's for

This is aimed at researchers, graduate students, and engineers working on visual SLAM, 3D reconstruction, or Transformer-based methods in computer vision. It is also useful for anyone surveying the field or looking for baseline papers to compare against.

Highlights

• Covers a broad range of SLAM-related tasks, from full SLAM pipelines to individual components like optical flow and depth estimation.
• Entries include direct links to papers and often to official code implementations, saving time on searching.
• Concise notes accompany each paper, highlighting the core idea or contribution.
• Actively encourages community contributions to keep the list comprehensive and up to date.