NVIDIA introduces MONAI cloud APIs, transforming medical imaging AI. These APIs, showcased at RSNA, expedite model development and data creation. Offering seamless integration and leveraging VISTA-3D for annotation, Auto3DSeg for training, and successful outcomes at MICCAI, the APIs empower solution providers. Industry leaders like Flywheel embrace these tools for rapid curation, labeling, and training, reducing costs and improving model quality. Additionally, annotation providers such as Redbrick AI leverage VISTA-3D, while MLOps platforms like Dataiku explore integration. NVIDIA MONAI cloud APIs revolutionize medical imaging AI, fostering efficiency, accuracy, and innovation in healthcare.
NVIDIA unveils MONAI cloud APIs, revolutionizing medical imaging AI at RSNA. These APIs accelerate AI model development and ground-truth data creation. Leveraging VISTA-3D and Auto3DSeg, they empower solution providers, as showcased at MICCAI. Industry leaders like Flywheel adopt these tools for efficient image curation, labeling, and training. Annotation providers such as Redbrick AI benefit from VISTA-3D, while MLOps platforms like Dataiku explore integration. NVIDIA MONAI cloud APIs usher in a new era of medical imaging AI, offering streamlined processes and enhanced accuracy, shaping the future of healthcare innovation.
The unveiling of NVIDIA MONAI cloud APIs at the RSNA annual meeting represents a significant leap forward in the field. These APIs are set to reshape the integration of AI into medical imaging offerings by leveraging pretrained foundational models and AI workflows specifically tailored for enterprises. Built upon the foundation of the open-source MONAI project, a collaborative effort between NVIDIA and King’s College London, these APIs aim to catalyze the creation of ground-truth data and the training of specialized AI models through a fully managed, cloud-based infrastructure.
The importance of medical imaging within healthcare cannot be overstated, as it constitutes around 90% of healthcare data. From aiding radiologists and clinicians in screenings, diagnoses, and interventions to assisting biopharma researchers in evaluating new drugs’ impact on clinical trial patients, medical imaging plays a pivotal role. Additionally, medical device manufacturers rely on it for real-time decision support.
Recognizing the scale and complexity of tasks across these domains, there’s a pressing need for a dedicated AI infrastructure specific to medical imaging—a robust platform capable of large-scale data management, ground-truth annotations, rapid model development, and seamless AI application deployment.
The introduction of NVIDIA MONAI cloud APIs addresses these challenges by enabling solution providers to seamlessly integrate AI into their medical imaging platforms. This integration empowers radiologists, researchers, and clinical trial teams with advanced tools to construct domain-specialized AI factories. These APIs are currently accessible in early access through the NVIDIA DGX Cloud AI supercomputing service.
Moreover, the NVIDIA MONAI cloud API has already found integration in Flywheel, a prominent medical imaging data and AI platform supporting end-to-end workflows for AI development. Other industry players, such as medical images annotation companies like RedBrick AI and machine learning operations (MLOps) platform providers like Dataiku, are also poised to incorporate NVIDIA MONAI cloud APIs into their offerings.
Efficient development of AI solutions in the medical imaging domain necessitates a robust foundation encompassing full-stack optimizations, scalable systems, and top-notch research, along with high-quality ground-truth data. The interactive annotation feature in the NVIDIA MONAI cloud APIs, powered by the VISTA-3D model, streamlines the creation of 3D segmentation masks essential for medical image analysis. Trained on an extensive dataset of annotated images from 3D CT scans spanning various diseases and body parts, the VISTA-3D model accelerates the annotation process while continuously improving its quality based on user feedback and new data.
Furthermore, the APIs include Auto3DSeg, simplifying the model development process by automating hyperparameter tuning and AI model selection for 3D segmentation tasks. NVIDIA researchers’ achievements at the MICCAI medical imaging conference, where they won four challenges using Auto3DSeg, underscore its efficacy in analyzing various medical scans, including 3D CT scans, brain MRIs, and 3D ultrasounds.
Industry leaders like Flywheel are embracing the NVIDIA MONAI cloud APIs to accelerate medical image curation, labeling analysis, and training, thereby reducing the cost of developing high-quality AI models. Similarly, annotation and viewer solution providers such as Redbrick AI are leveraging the VISTA-3D model to offer interactive cloud annotation for medical device customers, enhancing efficiency and accuracy in clinical applications.
Moreover, MLOps platform builders like Dataiku are exploring the integration of NVIDIA MONAI cloud APIs to simplify AI model creation for medical imaging applications, democratizing AI-powered workflows for both data and domain experts.
Overall, NVIDIA MONAI cloud APIs mark a transformative milestone in medical imaging AI. These APIs, demonstrated at RSNA and supported by VISTA-3D and Auto3DSeg successes at MICCAI, empower solution providers. Flywheel and Redbrick AI exemplify industry leaders benefiting from these tools, reducing costs and improving efficiency in image curation and training. As Dataiku explores integration, these APIs democratize AI-powered workflows, fostering collaboration between data and domain experts. NVIDIA MONAI cloud APIs offer unparalleled potential, driving advancements in healthcare, streamlining processes, and fueling innovation in medical imaging AI.
