Automatic Thyroid Ultrasound Image Segmentation Based on U-shaped Network
Jianrui Ding, Zichen Huang, Shi Mengdie, Chunping Ning · 2019
Automatic tumor segmentation of thyroid ultrasound image is quite challenging due to the poor image quality. Recently the U-shaped network, especially U-Net, has achieved good results in medical image segmentation. In this paper, we proposed a modified U-Net model (ReAgU-Net), which embedded the improved residual units into the skip connection among the encoding and decoding path and introduce the attention gate mechanism to multiply the weight feature maps obtained from shallow layers and deep layers. Also, a hyperparameter is introduced to combine Focal-Tversky Loss, Dice Loss and Cross-entropy Loss to jointly guide the model optimization process. The experimental results demonstrate that the proposed approach outperforms the other U-shaped models.