CP segmentation method based on cross-channel global information complementation

CN118982549APending Publication Date: 2024-11-19SOUTHWEAT UNIV OF SCI & TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410949530.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-16
Publication Date
2024-11-19

Smart Images

  • Figure CN118982549A_ABST
    Figure CN118982549A_ABST
Patent Text Reader

Abstract

The invention provides a CP (Colortal Polyps) segmentation method based on cross-channel global information complementation, and belongs to the technical field of medical image processing. The method is established on a PVT V2 framework, and comprises the following steps: firstly, introducing a decentralized constraint attention mechanism (CSA) module into a Transform branch of the PVT V2 framework, and enhancing detail semantic information extraction; secondly, a CNN Transform feature fusion (CTFF) module is designed on a CNN network feature fusion branch, a CNN attention mechanism gate module is designed on a Transform-CNN double parallel branch, and cross-channel fusion noise is inhibited through a parallel CNN; and finally, designing a loss function by adopting an In-depth supervision strategy, and carrying out weighted fusion on output of the three branches. The method not only improves the accuracy of polyp detection, but also is very suitable for clinical application due to extremely high calculation efficiency, so that the method has wide application potential in an actual medical environment. In addition, the light-weight model design also ensures the feasibility on resource-limited equipment, so that the system has important strategic significance in a modern medical system.
Need to check novelty before this filing date? Find Prior Art