Open yilinliu77 opened 4 years ago
Hello, We cannot provide the gt as it would be unfair to other evaluators. However, we provide with the source code a tool that allows you to compare the accuracy and completeness of your reconstructed mesh and it provides a visualization of the error map. -Neil
On Thu, Aug 8, 2019 at 6:34 PM whatseven notifications@github.com wrote:
We would like to use the ground truth of the model in the benchmark for further experiments. Such as visualizing the error map of the model or find the place with the lowest reconstrucbility. We will be very appreciated if you could share the ground truth of the benchmark to us.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/nmoehrle/uavmvs/issues/13?email_source=notifications&email_token=AAMILL4COLJE6N2FB42TMFTQDTCSJA5CNFSM4IKPW542YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HEJHFDQ, or mute the thread https://github.com/notifications/unsubscribe-auth/AAMILL27SIIXWLW2LSC2UE3QDTCSJANCNFSM4IKPW54Q .
-- Neil G. Smith, Ph.D. King Abdullah University of Science and Technology Research Scientist, GMSV neil.smith@kaust.edu.sa
How can we visualize the error map of our reconstructed mesh and the binary scene file(e.g. NY-1.bin)?
We would like to use the ground truth of the model in the benchmark for further experiments. Such as visualizing the error map of the model or find the place with the lowest reconstrucbility. We will be very appreciated if you could share the ground truth of the benchmark to us.