chaowentao / oacc-net Goto Github PK

PyTorch implementation of our paper "Occlusion-Aware Cost Constructor for Light Field Depth Estimation". [pdf]

• 2022-03-04: The camera ready version of our paper is available on arXiv.
• 2022-03-02: Our paper is accepted to CVPR 2022.
• 2022-02-28: Codes and models are uploaded.

• PyTorch 1.3.0, torchvision 0.4.1. The code is tested with python=3.6, cuda=9.0.
• A single GPU with cuda memory larger than 12 GB is required to reproduce the inference time reported in our paper.

• We used the HCI 4D LF benchmark for training and evaluation. Please refer to the benchmark website for details.

├──./datasets/
│    ├── training
│    │    ├── antinous
│    │    │    ├── gt_disp_lowres.pfm
│    │    │    ├── valid_mask.png
│    │    │    ├── input_Cam000.png
│    │    │    ├── input_Cam001.png
│    │    │    ├── ...
│    │    ├── boardgames
│    │    ├── ...
│    ├── validation
│    │    ├── backgammon
│    │    │    ├── gt_disp_lowres.pfm
│    │    │    ├── input_Cam000.png
│    │    │    ├── input_Cam001.png  
│    │    │    ├── ...
│    │    ├── boxes
│    |    ├── ...
│    ├── test
│    │    ├── bedroom
│    │    │    ├── input_Cam000.png
│    │    │    ├── input_Cam001.png  
│    │    │    ├── ...
│    │    ├── bicycle
│    |    ├── herbs
│    |    ├── origami

• Set the hyper-parameters in parse_args() if needed. We have provided our default settings in the realeased codes.
• Run train.py to perform network training.
• Checkpoint will be saved to ./log/.

• Place the input LFs into ./demo_input (see the attached examples).
• Run test.py to perform inference on each test scene.
• The result files (i.e., scene_name.pfm) will be saved to ./Results/.

• Perform inference on each scene separately to generate .pfm result files.
• Download groundtruth disparity images (i.e., gt_disp_lowres.pfm) and use the official evaluation toolkit to obtain quantitative results.

• Run test_inference_time.py to reproduce the inference time reported in our paper. Note that, the inference need to be performed on a GPU with a cuda memory larger than 12 GB.

Please refer to our supplemental material for additional quantitative and visual comparisons.

If you find this work helpful, please consider citing:

@inproceedings{OACC-Net,
  title     = {Occlusion-Aware Cost Constructor for Light Field Depth Estimation},
  author    = {Wang, Yingqian and Wang, Longguang and Liang, Zhengyu and Yang, Jungang and An, Wei and Guo, Yulan},
  booktitle = {IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
  year={2022}
}

Welcome to raise issues or email to [email protected] for any question regarding this work.