基于增强LetNet-5的非霍奇金淋巴瘤辅助诊断

针对主流网络模型在医学辅助诊断适用性低的问题,在现有LetNet-5模型的基础上,给出了增强LetNet-5模型用于非霍奇金淋巴瘤的智能辅助诊断方案。首先将获取到的数据集进行图像切分和归一化等预处理操作,然后使用深度学习框架KERAS搭建增强前后的LetNet-5模型,接着对增强前后的网络模型进行训练、预测和评估,最后对模型的泛化能力和稳定性进行验证。实验表明,增强LetNet-5模型相对于原始LetNet-5模型具有更高的识别精度、更好的稳定性和更快的模型收敛速度,为非霍奇金淋巴瘤的诊断提供科学性的指导并具有重要的临床价值。

Auxiliary Diagnosis of Non-Hodgkin's Lymphoma Based on Enhanced LetNet-5

In view of the low applicability of mainstream network model in medical aided diagnosis, based on the existing LetNet-5 network model, this paper presents an intelligent assistant diagnosis scheme for enhancing LetNet-5 model for non-Hodgkin"s lymphoma. Firstly, the obtained data set is subjected to pre-processing operations such as image segmentation and normalization. Then, the deep learning framework KERAS is used to build the LetNet-5 model before and after the enhancement, and then the network model before and after the enhancement is trained, predicted and evaluated. Finally, Verification of the generalization ability and stability of the model.Experiments show that the enhanced LetNet-5 model has higher recognition accuracy, better stability and faster model convergence speed than the original LetNet-5, it provides scientific guidance for the diagnosis of non-Hodgkin"s lymphoma and has important clinical value.