PATNet: Patch-to-pixel Attention-aware Transformer Network for RGB-D and RGB-T Salient Object Detection

Datasets

The training and testing datasets for RGBD and RGBT can be downloaded from the following links.

RGBD,RGBT

Pretraining Parameter

Backbone P2T weights: https://github.com/yuhuan-wu/P2T

RGBD: BaiduYun[8egc], Google Drive

RGBT: BaiduYun[7rjj], Google Drive

RGB-D Saliency map

BaiduYun[98vm], Google Drive

RGB-T Saliency map

BaiduYun[26rd], Google Drive

validation

The validation set is a randomly selected image from the test set BaiduYun[s2ne], Google Drive