SDETR: Attention-guided Salient Object Detection with Transformer

Most existing CNN-based salient object detection methods can identify fine-grained segmentation details like hair and animal fur, but often mispredict the salient object due to lack of global contextual information caused by locality convolution layers. The limited training data of the current SOD task add additional difficulty to capture the saliency information. In this paper, we propose a two-stage predict-refine SDETR model to leverage both benefits of transformer and CNN layers that can produce results with accurate saliency prediction and fine-grained local details. We also propose a novel pretrain dataset COCO SOD erase the overfitting problem caused by insufficient training data. Comprehensive experiments on five benchmark datasets demonstrate that the COCO SOD outperforms state-of-the-art approaches on four evaluation metrics, and our COCO SOD pretrain dataset can largely improve the model performance on DUTS, ECSSD, DUT, PASCAL-S datasets.