基于CT图像二维投影及深度学习的肺结核传染性判别方法

• Format: Patent
• Language: Chinese
• Published: 19.09.2025
• Subjects: CALCULATING; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

本发明涉及医学智能分析相关领域，公开了基于CT图像二维投影及深度学习的肺结核传染性判别方法，提供了一种基于二维投影的肺结核CT图像传染性分析方法，该方法不需要使用三维模型，且可以大量提升样本量，这大大缓解了深度学习模型的过拟合问题；提出双流的二维ResNet模型，该模型采用双主干网络设计，使用了残差模块提取特征，在二维投影的肺结核CT图像传染性判别任务上表现优秀；采用Z，X，Y三个方向的投影，且使用均值和标准差作为投影的属性，实验结果表明该做法是有效的；在测试阶段使用了三个方向的预测概率的均值，这是使得模型的最终预测结果更加具有鲁棒性。 The invention relates to the related field of medical intelligent analysis, discloses a pulmonary tuberculosis infectivity discrimination method based on CT image two-dimensional projection and deep learning, and provides a pulmonary tuberculosis CT image infectivity analysis method based on two-dimensional projection, which does not need to use a three-dimensional model, can greatly improve the sample size, and can improve the detection accuracy. The overfitting problem of the deep learning model is greatly relieved; a double-flow two-dimensional ResNet model is provided, the model adopts a double-backbone network design, features are extracted by using a residual module, and the model is excellent in infectious discrimination task of the tw