Machine and deep learning applications in medical imaging

Machine and deep learning have ushered in a transformative era in medical imaging, revolutionizing the way healthcare professionals diagnose and treat various conditions. These advanced technologies, powered by artificial intelligence, excel in analyzing vast amounts of medical imaging data with speed and precision. In the realm of diagnostics, machine and deep learning algorithms can detect subtle patterns and abnormalities in radiological images, such as ultrasound, MRI, and CT, assisting in early and accurate disease detection. Moreover, these systems can aid in image segmentation, enabling the delineation of specific structures or organs for treatment planning. A big focus of our group is developing AI-based solutions and integrating them into our image-guided navigated systems and training platforms.

Spine Reconstruction for Needle Insertion

Spinal needle insertions can be challenging, especially in deformed and diseased spine conditions. Although X-ray and computed tomography (CT) provides clear pictures for physicians to guide needles accurately in or near the spine, these imaging machines are sometimes not accessible, and their radiation may increase the risk of cancer if used for a long time. Ultrasound is a safe and accessible imaging modality, but a single ultrasound image only covers a small area of the spine. Our work enhances ultrasound images with AI (Artificial Intelligence) methods that filter bone surfaces from them. We use these images with 3D position tracking to reconstruct a 3D map of the spine that can be used for needle guidance at comparable accuracy to CT.