Semi-supervised prostate lesion grade identification model and system based on multi-task learning

The invention provides a semi-supervised prostate lesion grade identification model based on multi-task learning. The model structure sequentially comprises a U2-net neural network, a grey-scale map and original medical image conversion layer and an SE-ResNet34 neural network from an input end to an...

Full description

Saved in:
Bibliographic Details
Main Authors YI KUNMING, PENG JUNHAO, REN JUNJIE, TONG DALI
Format Patent
LanguageChinese
English
Published 10.05.2024
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:The invention provides a semi-supervised prostate lesion grade identification model based on multi-task learning. The model structure sequentially comprises a U2-net neural network, a grey-scale map and original medical image conversion layer and an SE-ResNet34 neural network from an input end to an output end. A U2-net neural network and an SE-ResNet-34 neural network are integrated, and when the neural networks are trained, fused semi-supervised multi-task learning is adopted. The depth of a network structure is greatly increased on the basis of a U-net network, and meanwhile, a U-block module in a U-net model is replaced with an RSU-L structure; and an SE mechanism is introduced on the basis of the ResNet34 neural network. Compared with an existing image segmentation recognition model, the technical scheme provided by the invention has the advantages that the segmentation and classification accuracy is obviously improved; the method has good generalization and robustness, can be used for transfer learning,
Bibliography:Application Number: CN202410149053