基于多图神经网络协同学习的显著性物体检测方法

In complex visual scene, the performance of existing deep convolutional neural network based methods of salient object detection still suffer from the loss of high-frequency visual information and global structure information of the object, which can be attributed to the weakness of convolutional neural network in capability of learning from the data in non-Euclidean space. To solve these problems, an end-to-end multiple graph neural networks collaborative learning framework is proposed, which realizes the cooperative learning process of salient edge features and salient region features. In this learning framework, this paper constructs a dynamic message enhancement graph convolution operator, which captures non-Euclidean space global context structure information by enhancing message transfer between different graph nodes and between different channels within the same graph node. Further, by introducing an attention perception fusion module, the complementary fusion of salient edge information and salient region information is realized, providing complementary clues for the two information mining processes. Finally, by explicitly encoding the salient edge information to guide the feature learning of salient regions, salient regions in complex scenes can be located more accurately. The experiments on four open benchmark datasets show that the proposed method has strong robustness and generalization ability, which make it superior to the current mainstream deep convolutional neural network based salient object detection methods.