The pytorch implementation of Multi-interactive Dual-decoder for RGBT Salient Object Detection
- We use VT5000-Train to train our network. All the datasets are available in https://github.com/lz118/RGBT-Salient-Object-Detection
- The pretrained model (VGG16) can be downloaded at https://pan.baidu.com/s/11lq3mUGRFP7TFvH9Eui14A [3513]
-
The trained models on RGB-T Dataset
https://pan.baidu.com/s/1Wj6bfi7lhp1KF5iCSVj0gQ [4zkx]
https://drive.google.com/file/d/11lU5TaRZMTXQ6QCbBLinG9iDvIUDrRP5/view?usp=sharing
-
The trained models on RGB-D Dataset
https://pan.baidu.com/s/1KlAKrVszQisG0bK1kiedzA [2ulc]
https://drive.google.com/file/d/1LKVn3iPDBI07DUBiirm4bk2-7yA3pTSM/view?usp=sharing
-
For RGB-T SOD, we provide the our saliency maps on VT821, VT1000 and VT5000-Test.
-
The saliency maps of all compared methods on VT821, VT1000 and VT5000-Test. https://pan.baidu.com/s/1s_pJ5qNJcQ8Q7ucusZHnRg [ax96]
-
For RGB-D SOD, we provide the our saliency maps on SIP, SSD，STERE，LFSD and DES.
-
The evalution toolbox is provided by https://github.com/jiwei0921/Saliency-Evaluation-Toolbox