Pioneering the Future of Medical Imaging: GE HealthCare’s Collaboration with NVIDIA on SonoSAMTrack, Large Medical Vision Foundation Model—In our quest to advance medical imaging, we at GE HealthCare are investing in the development of large vision foundation models, with our collaborative effort with NVIDIA serving as a cornerstone of this initiative. A key focus of our joint endeavor is the development of SonoSAMTrack, an innovative technology currently in the research and development phase. Developed using NVIDIA’s advanced AI and computing technologies, SonoSAMTrack is exploring new frontiers in medical diagnostics. While it remains experimental, this project is demonstrating its potential across various test scenarios. Our approach, which includes leveraging the NVIDIA TensorRT toolkit, is geared towards adapting to different medical imaging tasks with fewer data requirements, a critical aspect in the healthcare sector. The preliminary results from SonoSAMTrack have been promising, pointing to potential improvements in ultrasound imaging efficiency and accuracy. In its early tests, the model has achieved impressive accuracy scores, with over 90% on the Dice similarity coefficient scale, hinting at its high precision in image segmentation. Moreover, SonoSAMTrack is designed to be user-friendly, potentially requiring just 2-6 clicks for accurate segmentation. This ease of use could be a significant advancement in clinical settings. While SonoSAMTrack is not yet commercially available and continues to undergo development and regulatory approvals, its initial successes highlight our commitment to enhancing healthcare through AI innovation. Our collaboration with NVIDIA underscores this commitment, as we work together to push the boundaries of medical imaging. Stay tuned for more updates on the development of large vision models and our journey in AI-driven healthcare innovations. #gtc24 #gtc2024 #gehc #foundationmodels #ai #ultrasoundtechnology Jensen, Peter, David, Kimberly, Cao (Danica), Gopal, Erhan, Roland, Karley, Jan, Lynn, and Parminder (Parry)
Pioneering the Future of Medical Imaging: GE HealthCare’s Collaboration with NVIDIA on SonoSAMTrack GE HealthCare is committed to investing in innovative technologies that help tackle some of the industry’s biggest challenges. Our vision is to accelerate advancements in medical imaging by introducing foundational AI technologies, thereby empowering data scientists to expedite AI application development and eventually help clinicians and enhance patient care. By utilizing these versatile, generalist models, we aim to adapt more efficiently to new tasks and medical imaging modalities, often requiring far less labeled data compared to the traditional model retraining approach. This is particularly significant in the healthcare domain, for which data is especially time-consuming and costly to obtain. In a recent study conducted by GE HealthCare, its research project, SonoSAMTrack, showcased high performance across seven ultrasound datasets, encompassing a wide range of anatomies (adult heart and fetal head) and pathologies (breast lesions and musculoskeletal pathologies), as well as different scanning devices. This achievement was made possible through distillation and quantization techniques, utilizing the NVIDIA TensorRT software development kit and other capabilities for quantization-aware training. #gtc24 #gtc2024 #gehc #foundationmodels #ai #ultrasoundtechnology Roland Rott Taha Kass-Hout, MD, MS Karley Yoder Cao (Danica) Xiao Gopal Avinash Ph.D Jan Beger Erhan Bas
Kristi Henderson
7mo
Taha Kass-Hout, MD, MS Exciting update. Looking forward to hearing more as the partnership progresses!
David Niewolny
7mo
Great work by Parminder Bhatia and team! So much more to come.
Vice President, Global and Strategic Alliances; Tech, Services, Systems Integrators, and Bio-Pharma Alliances Professional; Builder of Ecosystems
7moAwesome. Looking forward to elevating this partnership!