MINIMA: Modality Invariant Image Matching
This Space is derived from Image Matching WebUI. We are grateful to the authors for their contribution of the source code.
Here we provide our MINIMA-model in our paper for test and comparison, and this project is undergoing continuous enhancement.
Special thanks to Hugging Face for providing the ZeroGPU for this Space!
Matching Model
48 2048
64 2048
0 1
10 10000
0 1
0.1 1
RANSAC Method
0 20
0 1
0 100000
Examples (click one of the images below to Run Match).
| Image 0 | Image 1 | Match threshold | Max features | Keypoint threshold | Matching Model | RANSAC Method | Ransac Reproj threshold | Ransac Confidence | Ransac Iterations |
|---|
Pages:
Algo. | Conference | Code | Paper | Project |
|---|---|---|---|---|
LightGlue | ArXiv 2024 | https://github.com/cvg/LightGlue | https://arxiv.org/abs/2412.19412 | https://parskatt.github.io/RoMa |
Camera Model
Matching Model
100 10000
0 1
0.01 12
0.01 12
0.01 1
1 100
Scene Graph
Global features
1 100