What does a scanner see?: Visualising anatomy in 3D

16:00-16:30, Lecture Theatre ("Raked")

Visualising biological information is challenging at the best of times. At axial3D, we accelerate that understanding by providing Machine Learning (ML) backed annotations of those images. We develop these algorithms and the tools to visualise their output in-house at our Belfast office. In this talk, we will outline our ML training pipeline and show how we are harnessing the power of the cloud to provide an intuitive web-based platform for the annotation of medical images in a secure and scalable way.

Currently, there is no standardised pipeline for the creation, processing and validation of image segmentation for medical 3D printing. The features and functions exist in a disparate array of web and desktop-based software. We will show how we are bringing the required functionality together in a web-based tool. This tool can handle any step of the process from medical image transfer to printing the 3D object. We will show how we capture and curate training data, as well as train the actual ML pipeline.

Attendees can expect to learn practically how to ingest data into a machine learning platform. How to visualise that data and securely allow others to do so. How to represent potential mistakes in image segmentation tools and produce a web-based platform for the curation of ML data.