no code implementations • 21 Mar 2022 • Xian Fang, Jinshao Zhu, Xiuli Shao, Hongpeng Wang
The features of the intermediate process are first fused by the features of different layers, and then processed by several transformers in multiple groups, which not only makes the size of the features of each scale unified and interrelated, but also achieves the effect of sharing the weight of the features within the group.
no code implementations • 21 Oct 2021 • Xian Fang, Jinchao Zhu, Xiuli Shao, Hongpeng Wang
Currently, existing salient object detection methods based on convolutional neural networks commonly resort to constructing discriminative networks to aggregate high level and low level features.
no code implementations • 16 Sep 2021 • Xian Fang, Jinchao Zhu, Ruixun Zhang, Xiuli Shao, Hongpeng Wang
The adjacent interactive aggregation module (AIAM) gradually integrates the neighbor features of high, middle and low levels.
no code implementations • 13 Aug 2021 • Jinchao Zhu, XiaoYu Zhang, Xian Fang, Feng Dong, Qiu Yu
Then, a modal-adaptive gate unit (MGU) is proposed to suppress the invalid information and transfer the effective modal features to the recoding mixer and the hybrid branch decoder.
no code implementations • 27 Jul 2021 • Jinchao Zhu, XiaoYu Zhang, Xian Fang, Feng Dong, Li Yuehua, Junnan Liu
Effective fusion of different types of features is the key to salient object detection.