Home

Awesome

StereoNeRF<br><sub>Official PyTorch Implementation</sub>

Generalizable Novel-View Synthesis using a Stereo Camera<br> Haechan Lee*, Wonjoon Jin*, Seung-Hwan Baek, Sunghyun Cho<br>

News

Abstact

In this paper, we propose the first generalizable view synthesis approach that specifically targets multi-view stereo-camera images. Since recent stereo matching has demonstrated accurate geometry prediction, we introduce stereo matching into novel-view synthesis for high-quality geometry reconstruction. To this end, this paper proposes a novel framework, dubbed StereoNeRF, which integrates stereo matching into a NeRF-based generalizable view synthesis approach. StereoNeRF is equipped with three key components to effectively exploit stereo matching in novel-view synthesis: a stereo feature extractor, a depth-guided plane-sweeping, and a stereo depth loss. Moreover, we propose the StereoNVS dataset, the first multi-view dataset of stereo-camera images, encompassing a wide variety of both real and synthetic scenes. Our experimental results demonstrate that StereoNeRF surpasses previous approaches in generalizable view synthesis.

Installation & Dependency

Tested with Docker image (pytorch/pytorch:1.9.1-cuda11.1-cudnn8-devel).

To install python library dependencies, run the code below:

pip install -r requirements.txt

If you want to reproduce the results of StereoNeRF:

You can download the pretrained weights and rendered StereoNVS-Real images of StereoNeRF from our Google Drive Link.

(Our code is heavily based on GeoNeRF).

Train

To train the model, run:

python run_stereo_nerf.py --config configs/config_general_stereo_real.txt

Inference

Set the ckpt_file path properly at line 667 in 'run_stereo_nerf.py', then run:

python run_stereo_nerf.py --config configs/config_general_stereo_real.txt --eval

Metric

To calculate the scores, run:

python calc_score.py --pred /path/to/evaluation/folder/in/prediction

License of StereoNVS Dataset

Contact

If you want to contact the author, email to: gocks8@gmail.com