AI segmentation of anatomical structures with apoQlar
apoQlar is the creator of VSI HoloMedicine® platform that leverages the Microsoft HoloLens 2 hardware to transform medical images, clinical workflows, and medical education into a 3D mixed reality environment.
The innovative approach apoQlar is using relies on processing the real data of the patients, to show the exact structures that can be used by the surgeons in surgical planning activities or in patient education to explain the planned surgery. As the exact shape, size, and position of anatomical structure, as well as presence of different malformations vary from human to human, apoQlar needed to take advantage of the power of advanced AI models to make their vision come true.
Together with the experts from apoQlar and with the help of doctors specialized in different fields of medicine we developed a set of AI models for the segmentation of various anatomical structures based on Magnetic Resonance Imaging (MRI). The models included structures such as bones, vessels, or structures within the brain such as the ventricles.
We have built the AI models architecture based on the state-of-the-art neural network architectures, including the variations of the U-Net architecture, achieving the best results in the field.
Applying AI in to support medical devies, for purposes other than research, requires medical certification. All major regulations require robust documentation and reproducibility of the AI models, with regards to areas such as data management, model training and tuning, model evaluation and model deployment. It’s important to develop AI models using frameworks and methods allowing to store e.g. dataset characteristics, training procedure and results or achieved metrics. Therefore, our models are built based on MS Azure Machine Learning and MS InnerEye.
How can we help you? The fields marked as required below will help us to process your request and forward it accordingly. You can expect a response within one to two business days.