PANet

Official code repository for paper Parallax-Aware Network for Light Field Salient Object Detection (SPL2024).

Overall Architecture

Overall architecture of the proposed PANet.

Details of the Parallax Processing Module (PPM).

Inference

Download the pretrained model at Baidu Driver (PSW: xf97).

Results

Quantitative Results. “N/T” denotes unavailable results. ↑/↓ means that a larger/smaller value is better. The top two results are highlighted in red and blue, respectively.

Download

The saliency maps can be download from Baidu Driver (PSW: kxxi).

Citation

Please cite the following paper if you use this repository in your research:

@article{yuan2024parallax,
  title={Parallax-Aware Network for Light Field Salient Object Detection},
  author={Yuan, Bo and Jiang, Yao and Fu, Keren and Zhao, Qijun},
  journal={IEEE Signal Processing Letters},
  year={2024},
  publisher={IEEE}
}

Any questions regarding this work can contact yuanboera@qq.com.