This repository contains code to compute depth from a single image. It accompanies our paper:

Improving Monocular Depth Estimation by Leveraging Structural Awareness and Complementary Datasets
Tian Chen⋆, Shijie An⋆, Yuan Zhang, Chongyang Ma, Huayan Wang, Xiaoyan Guo, and Wen Zheng

1. Download the model weights (full model training on NYUv2 dataset) SANet-NYUv2.pth(password:x55o) and place the file in the pretrained folder.

2. Set up dependencies:

   conda install pytorch torchvision opencv

   The code was tested with Python 3.6, PyTorch 1.1.0, and OpenCV 4.1.2.

1. Place the pretrained model in the folder pretrained.

2. Run the model:

   python test.py --cuda

(1) Our network architecture

(2) Our result compare on NYUv2 dataset

Please cite our paper if you use this code or any of the models:

@article{Tian2020,
	author    = {Tian Chen⋆, Shijie An⋆, Yuan Zhang, Chongyang Ma, Huayan Wang, Xiaoyan Guo, and Wen Zheng},
	title     = {Improving Monocular Depth Estimation by Leveraging Structural Awareness and Complementary Datasets},
	journal   = {},
	year      = {2020},
}

MIT License