Hybrid deep feature and machine learning framework for classification of thyroid nodules in ultrasound images
-
By
-
Dingnan Zhang
-
Bo Li
-
Hao Ju
-
Tingxue Li
-
Yanzhu Zhang
-
May 25, 2026
Objective:
To enhance the reliability of ultrasound-based thyroid nodule assessment through an optimized computer-aided diagnosis (CAD) framework.
Key Findings:
- The proposed framework achieves discrimination between benign and malignant thyroid nodules compared to conventional approaches.
- It demonstrates improved robustness across variations in ultrasound appearance.
- The model maintains stable performance without extensive parameter tuning.
Interpretation:
Limitations:
- The study does not address the performance of the model in real-world clinical settings.
- Limited exploration of hybrid models in pathology-linked thyroid ultrasound classification.
Conclusion:
The study presents a model that integrates deep learning and traditional machine learning for thyroid nodule classification.
