Medical image segmentation is an important part of clinical diagnosis.Accurate segmentation is of great significance to clinical treatment.In order to solve the problems of poor robustness and weak anti-noise ability of existing image segmentation algorithms,the study proposes a CT image segmentation algorithm based on deep learning.In the algorithm,the U-Net network structure is improved,a batch standardization layer is added to improve the robustness of the network model,and an attention mechanism is introduced to focus on specific features.Meanwhile,the cross entropy loss function is used to reduce segmentation deficiency and segmentation leakage,improve the segmentation accuracy,and train the segmentation network on the basis of data preprocessing to obtain accurate segmentation results.Compared with other algorithms,the Dice coefficient of commonly used evaluation criteria reaches 0.94,and the method has strong robustness.It can accurately segment lung organs in CT images,which is helpful to assist the diagnosis of lung diseases.
关键词
医学图像分割/U-Net/CT图像分割算法/深度学习
Key words
Medical image segmentation/U-Net/CT image segmentation algorithm/Deep learning
维普
万方数据