yukiumi13
commented
4 days ago Hello Soboliev,
Thank you for your question.
It takes similar time to NeuS, i.e., 3h per object on average. We have tried employing several acceleration techniques like hashgrid in instan-ngp, which can reduce the time within a few minutes. However, the reconstruction quality will be worse with some floating artifacts. Adding disparity loss in NeuralAngelo, fascinating the opacity to be 0 or 1, helps clean the floatings. But its quality is still worse than the original version.
发件人: Soboliev Oleksandr @.> 发送时间: Monday, October 14, 2024 5:12:01 PM 收件人: yukiumi13/GNeRP @.> 抄送: Subscribed @.***> 主题: [yukiumi13/GNeRP] How long does it take? (Issue #1)
Hi, thanks a lot for your work in this research, I read the paper, but haven't found any information, of the duration of reconstruction for your dataset, and how does it depend on size of the input data. That would be really helpful
― Reply to this email directly, view it on GitHubhttps://github.com/yukiumi13/GNeRP/issues/1, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AKQSOBD34CPTSOJPVO4CB6LZ3ODGDAVCNFSM6AAAAABP4RC746VHI2DSMVQWIX3LMV43ASLTON2WKOZSGU4DKMZVGUYTGNA. You are receiving this because you are subscribed to this thread.Message ID: @.***>
Hi, thanks a lot for your work in this research, I read the paper, but haven't found any information, of the duration of reconstruction for your dataset, and how does it depend on size of the input data. That would be really helpful