Tomosuite™
UX/UI, Iconography, Brand Design
Project in BWH's Department of Emergency Radiology
What is Tomosuite™?
Current medical imaging data is often clinician-oriented and is not easily amenable to machine learning. Furthermore, AI driven research in radiology can be inaccessible due to immense human and computer resource requirements. Tomosuite™ is an AI platform developed at Brigham and Women's Hospital that provides a fast and intuitive workbench for radiologists and data scientists to collaborate on deep learning research projects.
My Role Overview
As a UX/UI + Branding Design Intern with Brigham & Women's Hospital in the Department of Emergency Radiology, I designed and implemented a user interface and brand identity with custom letterforms for the Tomosuite™ platform, which consists of Studio, Central, Cohort+, and SysAdmin, as shown below. I collaborated with BWH Emergency Radiology physicians to develop interfaces that respond to their needs with modern design philosophies. I optimized the efficiency of department and radiologist research and development of decision support tools through UX/UI improvements in existing management and AI model inference technologies.
Tomosuite™ Central
Tomosuite™ Central is a web application that gives clinicians and data scientists the ability to oversee all stages of the ML development and deployment pipeline. Data, projects, resources, and even AI models can be studied and managed in this centralized application. Below are two prototype screens I designed for Tomosuite™ Central.
Tomosuite™ Studio
Tomosuite™ Studio is a clinician-oriented annotation and visualization platform that is designed to function similar to a clinical PACS viewer. It enables clinicians to efficiently annotate and review images which are used in all stages of the ML pipeline, and even includes an AI-powered tool to help them annotate faster. The image below is what the interface looked like when I first onboarded, along with some preliminary notes I took.
Revamped UI for Optimized User Flow
Radiologists' time is very limited, and my goal was to optimize the Tomosuite™ Studio platform so that they could get their job done quickly. That involved upgrading iconography for faster identification of function, improved organization to direct workflow components, and tightening the UI layout so that a user can assess information without being too overwhelmed.
Results
These improvements contributed to the efficiency of the platform, enabling radiologists to collaborate despite the many other competing duties they have. The advances this platform has made are documented in its arXiv publication.