mvsnerf
mvsnerf copied to clipboard
apchenstu
[ICCV 2021] Our work presents a novel neural rendering approach that can efficiently reconstruct geometric and neural radiance fields for view synthesis.
I notice that the batchsize can only be one.(One scene). Can it train multiple scenarios at once?
I am a little bit confused about the get_ndc_coordinate() Is That becasue In the supplymenty material of Nerf. The NDC convert aims to the **rays_o** and **rays_d**. And in this...
Thanks for the amazing work. When I download the dtu_training.rar files, the depth map resolution is 128x160, while in dtu_example, the depth map resolution is 1200x1600. Where should I get...
(mvsnerf) wulijun@dl-SYS-4028GR-TR:~/xwj/mvsnerf-main$ CUDA_VISIBLE_DEVICES=0,1 python train_mvs_nerf_pl.py --expname training_dtu --num_epochs 6 --N_samples 128 --use_viewdirs --batch_size 1 --dataset_name dtu --datadir /hdata/wulijun/xwj/mvsnerf-main/datasets/dtu --N_vis 6 Found ckpts [] GPU available: True, used: True TPU available:...
Hi , is there anyone who can help solve this problem Found ckpts [] GPU available: True, used: True TPU available: False, using: 0 TPU cores ==> image down scale:...
I saw you compare the depth estimation with pixel nerf and ibrnet. I also interested in depth estimation with other nerfs. Did you compare with others? Can you share some...
Can somebody tell me what the relationship is between these two data sets? And how to follow with the example's folder structure.
when I run the renderer.py ,How should this problem be solved --------------------------ValueError Traceback (most recent call last) Input In [10], in () 108 rgb_rays = rgb_rays[H_crop:-H_crop,W_crop:-W_crop] 110 psnr.append( mse2psnr(np.mean((rgb_rays-img)**2))) -->...
