Lattice curved surface array based collaborative fusion for intelligent diagnosis

作者：Peng, Weimin^*; Deng, Huifang; Chen, Aihong; Huang, Wenyuan

来源：Information Fusion, 2023, 93: 77-84.

DOI：10.1016/j.inffus.2022.12.021

• Lattice curved surface array
• Multi-source medical examination data
• Collaborative fusion
• Reliability evaluation
• Precise pneumonia diagnosis

A new concept of lattice curved surface array is proposed in this paper to represent multi-source medical examination data for getting precise conclusions of intelligent pneumonia diagnosis through reliable collaborative fusion. To this end, the quantum phase based lattice curved surfaces of all source datasets are constructed to form a lattice curved surface array which provides the precise division and fusion space for all entity nodes. Based on the horizontal category degrees and vertical collaborative degrees between entity nodes, the constructed lattice curved surface array is segmented into different subset arrays, and the entity nodes in each subset array are fused into a single object entity node according to the weights of their included part nodes. The experimental data evaluation shows that, compared with the homogeneous fusion methods, this new concept based collaborative fusion method is impressively effective with a preciseness of 95.52% achieved in the reliable fusion results based pneumonia diagnosis.