Related Experiment Video
Updated: Jun 10, 2025
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
Published on: April 21, 2023
Enhanced pediatric thyroid ultrasound image segmentation using DC-Contrast U-Net
Bo Peng1,2, Wu Lin3, Wenjun Zhou4,3
1Ultrasound in Cardiac Electrophysiology and Biomechanics Key Laboratory of Sichuan Province, Sichuan Provincial People's Hospital, University of Electronic Science and Technology of China, Chengdu, 611731, China. bpeng1@vip.163.com.
Insights
A new AI model, DC-Contrast U-Net, improves pediatric thyroid ultrasound segmentation accuracy and speed. This advanced deep learning approach offers a more objective and efficient method for early thyroid screening in children.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Pediatric Endocrinology
Background:
- Early thyroid gland screening is crucial, but current methods like palpation and ultrasound have limitations in pediatric populations.
- Accurate thyroid gland localization and size determination in children via ultrasound are challenging due to operator experience and subjectivity.
- Existing ultrasound techniques suffer from inter-observer variability and depend heavily on operator expertise.
Purpose of the Study:
- To develop a novel, computationally efficient deep learning model for segmenting pediatric thyroid ultrasound images.
- To improve the accuracy and reduce the complexity of medical image segmentation for pediatric thyroid analysis.
- To address the limitations of subjective and experience-dependent ultrasound assessments in pediatric thyroid screening.
Main Methods:
- A novel U-Net-based network, termed DC-Contrast U-Net, was designed for enhanced texture information extraction from pediatric thyroid ultrasound images.
- The model focuses on reducing computational complexity and the number of parameters while maintaining high segmentation performance.
- Comparative analysis was performed against other U-Net-related segmentation models.
Main Results:
- The proposed DC-Contrast U-Net model demonstrated superior segmentation accuracy compared to existing U-Net-related models.
- The network achieved improved inference speed, indicating greater computational efficiency.
- The model shows potential for accurate and objective thyroid gland segmentation in pediatric patients.
Conclusions:
- DC-Contrast U-Net offers a promising solution for accurate and efficient pediatric thyroid ultrasound image segmentation.
- The model's improved performance and reduced complexity make it suitable for clinical applications and deployment on medical edge devices.
- This deep learning approach can enhance early thyroid screening in children by providing more reliable and objective diagnostic information.
Abstract:
Early screening methods for the thyroid gland include palpation and imaging. Although palpation is relatively simple, its effectiveness in detecting early clinical signs of the thyroid gland may be limited, especially in children, due to the shorter thyroid growth time. Therefore, this constitutes a crucial foundational work. However, accurately determining the location and size of the thyroid gland in children is a challenging task. Accuracy depends on the experience of the ultrasound operator in current clinical practice, leading to subjective results. Even among experts, there is poor agreement on thyroid identification. In addition, the effective use of ultrasound machines also relies on the experience of the ultrasound operator in current clinical practice. In order to extract sufficient texture information from pediatric thyroid ultrasound images while reducing the computational complexity and number of parameters, this paper designs a novel U-Net-based network called DC-Contrast U-Net, which aims to achieve better segmentation performance with lower complexity in medical image segmentation. The results show that compared with other U-Net-related segmentation models, the proposed DC-Contrast U-Net model achieves higher segmentation accuracy while improving the inference speed, making it a promising candidate for deployment in medical edge devices in clinical applications in the future.
