OP02.06: Application of knowledge distillation technology for fine segmentation of three‐vessel and trachea views in fetal echocardiographic images
Year: 2024
Type: article
Abstract: To explore fetal heart ultrasound image segmentation network model based on knowledge distillation technology in the fine segmentation of fetal echocardiographic image at three-vessel and trachea (3VT) views. Total 1300 two-dimensional grayscale ultrasound images of fetal heart at 3VT views were collected and divided into training, validation, and testing sets. The training and validation sets were used to construct the auxiliary diagnostic network models, and the test sets were used to test different network models (U-Net, DeepLabv3+). The 3VT views were collected and annotated by an experienced doctor as the gold standard. The intersection over union (IoU), pixel accuracy (PA) and Dice coefficient (Dice) were used as the three indexes to evaluate the segmentation accuracy, and the diagnostic efficiency of the training model was evaluated. This model and the most commonly used segmentation models were identified, and the results were compared. 101 images were randomly selected and assigned to junior doctors, AI and junior doctors to assist AI interpretation. Bland-Altman images were drawn to evaluate their consistency with the gold standard, and the results of the three were compared. The training model of knowledge distillation algorithm achieved better results than other models on all evaluation indexes, and the average IoU, PA and Dice were 68.6%, 81.4% and 81.3%, respectively. Compared with the U-Net model and DeepLabv3+ model, more accurate segmentation boundaries were obtained by this model, and the quantitative evaluation indexes were improved. With the aid of the model, the accuracy of the diagnosis of junior doctors was improved. The knowledge distillation algorithm training segmentation model can identify the anatomical structure of the fetal heart at VT view and the recognition result is obviously better than that of the related methods, and can improve the accuracy of the image recognition of the doctors with less experience.
DOI: 10.1002/uog.27888
Institutions: Sir Run Run Shaw Hospital, Zhejiang University
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W4402360787