Salient object detection on hyperspectral images using features learned from unsupervised segmentation task
This pytorch code was implemented by Nevrez Imamoglu and Guanqun Ding for the following paper:
Nevrez Imamoglu, Guanqun Ding, Yuming Fang, Asako Kanezaki, Toru Kouyama, Ryosuke Nakamura, "Salient object detection on hyperspectral images using features learned from unsupervised segmentation task", in IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), 2019. (pdf)
This code was based on the Asako Kanezaki's awesome work: Asako Kanezaki, "Unsupervised Image Segmentation by Backpropagation", in IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)*, 2018. (pdf)
python3.6, pytorch, scipy==1.2.0,opencv-python, h5py, scikit-image==0.14.0
install torch:
sudo pip3 install torch torchvision
install opencv:
sudo pip3 install opencv-python
- put your HS data into folder
./HS_images
, you can download our HS-SOD Dataset from here, (pdf, project, dataset) - run the script on bash
sh run.sh
- after iteration, you can find the training results in folder './HS_Results', the final saliency maps can be found in './HS_Results/sal_result'
If this code is useful for your research, please cite the following paper:
[1] Nevrez Imamoglu, Guanqun Ding, Yuming Fang, Asako Kanezaki, Toru Kouyama, Ryosuke Nakamura, "Salient object detection on hyperspectral images using features learned from unsupervised segmentation task", in IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), 2019. (pdf)
[2] Asako Kanezaki, "Unsupervised Image Segmentation by Backpropagation", in IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)*, 2018. (pdf)
[3] Nevrez Imamoglu, Yu Oishi, Xiaoqiang Zhang, Guanqun Ding, Yuming Fang, Toru Kouyama, Ryosuke Nakamura, "Hyperspectral Image Dataset for Benchmarking on Salient Object Detection", 10th International Conference on Quality of Multimedia Experience (QoMEX), Sardinia, Italy, May 29 - June 1, 2018. (pdf, project, dataset)