360-SSOD / download

Requirements

• Linux
• Python 3.5+ (Say goodbye to Python2)
• PyTorch 1.0+ or PyTorch-nightly
• CUDA 9.0+
• NCCL 2+
• GCC 4.9+
• mmcv

We have tested the following versions of OS and softwares:

• OS: Ubuntu 16.04/18.04 and CentOS 7.2
• CUDA: 9.0/9.2/10.0
• NCCL: 2.1.15/2.2.13/2.3.7/2.4.2
• GCC: 4.9/5.3/5.4/7.3

Install mmdetection

1. Create a conda virtual environment and activate it. Then install Cython.

2. Install PyTorch stable or nightly and torchvision following the official instructions.

3. Clone the mmdetection repository.

git clone https://github.com/360-SSOD/download.git

4. RPN models are available in the Model

Results

360-SOD&360-SSOD Tencent Cloud Link

Datasets

360-SSOD : Tencent Cloud Link

Usage

1. Please download our pretrained model.
   Put this model in ./model/ and ./reimg/.

2. Run: test.py

Evaluation Method

Saliency Evaluation Method : The code can be found in Ming-Ming Cheng and Deng-Ping Fan.

Citation:

@article{ma2020stage,
  title={Stage-wise salient object detection in 360° omnidirectional image via object-level semantical saliency ranking},
  author={Ma, Guangxiao and Li, Shuai and Chen, Chenglizhao and Hao, Aimin and Qin, Hong},
  journal={IEEE Transactions on Visualization and Computer Graphics},
  volume={26},
  number={12},
  pages={3535--3545},
  year={2020},
  publisher={IEEE}
}

Contacts

[e-mail]