The second application addresses the challenge of intervertebral disc (IVD) segmentation in magnetic resonance imaging (MRI). IVDs are structures located between vertebrae, responsible for spinal flexibility and load distribution. However, training DL models for accurate segmentation often requires large amounts of labeled data, a significant bottleneck in medical imaging. To overcome this, Francesca developed a domain adaptation model using contrastive learning, enabling the model to learn from unlabeled data by performing pretext tasks on intensity features. This approach reduces dependency on expert-labeled data and improves the model's ability to generalize across different datasets, handling variations caused by different MRI devices and medical conditions. The model demonstrated robust performance, advancing the adaptability and efficiency of DL tools in medical imaging. The last application focuses on carpal tunnel syndrome (CTS) and its evaluation using ultrasound (US) imaging. CTS, caused by the compression of the median nerve in the wrist, leads to pain, numbness, and weakness in the hand. Diagnosing CTS typically involves patient history, physical examination, and imaging studies. US imaging allows precise visualization of the median nerve structure. Francesca developed a DL framework to automatically segment the median nerve at the carpal tunnel inlet, enabling consistent and objective measurement of the nerve cross-sectional area, a key metric in diagnosing and monitoring CTS. By reducing variability between clinicians' assessments, her framework improves the accuracy and reliability of CTS diagnosis and progression monitoring, ultimately improving patient outcomes. Francesca’s work highlights the potential of DL to revolutionize medical imaging. By addressing specific diagnostic challenges, her contributions pave the way for more consistent, efficient, and accurate assessments in clinical practice. Her innovative approaches mark a significant advancement in medical diagnostics, establishing a novel paradigm of precision, reliability, and standardization in patient evaluation and treatment planning. 39 Francesca Pia Villani Computer Vision News Computer Vision News
RkJQdWJsaXNoZXIy NTc3NzU=