search

zju3dv / NeuralRecon

Code for "NeuralRecon: Real-Time Coherent 3D Reconstruction from Monocular Video", CVPR 2021 oral
https://zju3dv.github.io/neuralrecon/
Apache License 2.0
2.07k stars 299 forks source link

Custom Dataset #22

Closed jeffin07 closed 3 years ago

jeffin07 commented 3 years ago

Hi, First of all great work. I would like to know if I want to test this method using my own dataset, say I have a small video and u want to try it out what are the steps that I need to follow before I run the inference. I found that you follow the scannet data structure, so can you tell me the steps for converting a custom dataset that will be able to run this method

JiamingSuen commented 3 years ago

https://github.com/zju3dv/NeuralRecon/blob/master/DEMO.md

jeffin07 commented 3 years ago

@JiamingSuen Thank you for the reply. So in order to run with custom data, it requires an iOS environment. Is that correct ? Is there a way if we can run it using a custom dataset without iOS, like can we convert a video into Scannet structure ?

JiamingSuen commented 3 years ago

Yes, an iOS device is required. Similar to this reply, NeuralRecon requires camera poses with metric scale as input.