search

prstrive / UniMVSNet

[CVPR 2022] Rethinking Depth Estimation for Multi-View Stereo: A Unified Representation
MIT License
228 stars 12 forks source link

about F-score on tanksandtemples #11

Closed CC9310 closed 10 months ago

CC9310 commented 2 years ago

I have a problem on the tanks dataset: for all 8 scenes, the private testing result is 79.61 65.66 54.16 59.51 62.24 60.90 57.61 54.59

i follow the author's resconstruction configuration, while there is a large margin compare with the author's result, whether i miss something important?

prstrive commented 2 years ago

I have limited ideas about the reason only from this succinct statement. Maybe there are something wrong on environment or fusion configuration?

CC9310 commented 2 years ago

@prstrive thanks for your respond, i will recheck the depth estimation process. 🙂