Figure.1 The overall architecture of the proposed E2Net model, which is published in Applied Intlligence 🎆. The paper can be downloaded from here[code:NEPU].
Python v3.6, Pytorch 0.4.0+, Cuda 10.0, TensorboardX 2.0, opencv-python
Download the train data from here[code:NEPU], test data from here[code:NEPU], test_in_train data from here[code:NEPU]. Then put them under the following directory:
-Dataset\
-train\
-test\
-VT821\
-VT1000\
-VT5000_test\
-test_in_train\
- Training the E2Net
Please download the released code and the data set, then:
run python train.py
- Testing the E2Net
Please download the trained weights from here[code:NEPU], and put it in './pre' folder, then:
run python test.py
Then the test maps will be saved to './Salmaps/'
- Evaluate the result maps
You can evaluate the result maps using the tool from here[code:NEPU], thanks for Dengpin Fan.
- Qualitative comparison
Figure.2 Qualitative comparison of our proposed method with some SOTA methods.
- Quantitative comparison
Table.1 Quantitative comparison with some SOTA models on three public RGB-T benchmark datasets.
- Salmaps
The salmaps of the above datasets can be download from here [code:NEPU]
If you have any questions, feel free to contact us via wuranwan2020@sina.com (Ranwan Wu). For more related work, you can also visit wuranwan.