Haotong LIN

Yes, if everything goes well, I will release this part of the code before SIGGRAPH Asia 2023

Although I haven't conducted that particular experiment yet, my experience with other datasets suggests that training a model with full views (21 views for ZJU-MoCap) and an input ratio of...

The zjumocap dataset is captured with 21 industrial cameras (2048x2048). We resize the images to 1024*1024. I think the estimation of smpl keypoints under different resolutions will not affect the...

1. Bckground.ply is the SFM sparse point cloud which is computed during calibration. 2. Outdoor dataset does not needs human mask information. To obtain the 3d bbox, you can follow...

If you're interested in processing static scenes captured by a single camera, you might find it helpful to follow the NeRF tutorial for converting the data to either LLFF format...

Hi, thanks for your attention! The 3D bounding box can be defined as very rough hyper-parameter, because our algorithm does not require an accurate 3D bounding box. If the foreground...

Thank you for your attention! We will release the interactive rendering, other scripts and a guide on how to run enerf on custom data soon. (Expected in the weekend.)

> I also want to confirm that ENeRF can generate a frame at **any** time, right? For example, synthesize a frame at a novel time t1.5 that is not in...

> Thank you very much for the plan of releasing such a guide. May I ask whether you can include the instructions to run the "compositional ENeRF" in the guide...

> We have released the dataloader for ZJU-MoCap. Thanks for your attention!