Submission details
| Method name | LoFTR [local feature challenge] |
| Dataset | aachenv11 |
| Submission date | March 26, 2021, 7:39 a.m. |
| Last modified | April 9, 2021, 1:36 a.m. |
| Publication url | https://zju3dv.github.io/loftr/ |
| Code url | https://github.com/zju3dv/LoFTR |
| Other info | LoFTR + coarse-level matching with dual-softmax (LoFTR-DS). LoFTR trained on MegaDepth. Using the official challenge pipeline. |
Detailed results for all conditions
| Method | day | night |
|---|---|---|
| LoFTR [local feature challenge] | 0.0 / 0.0 / 0.0 | 72.8 / 88.5 / 99.0 |