Computer Vision for Object Detection in Healthcare

Instruments left inside a patient at the end of an operation is a rare but preventable surgical error. The standard practice to deal with it is to manually count all instruments and sponges at the start and conclusion of a surgical procedure. Unfortunately, manual counting is not a fool-proof method and is subject to human error.
To increase the accuracy of counting, the project trained a computer vision model to identify surgical instruments. The developed system counts the instruments it detects in the operation tray, displays their names and records instruments removed from the tray as being in-use. 

UCL Computer Science BSc Project 2019

Library for building SMART on FHIR applications

Many hospitals and healthcare providers lack the ability to exchange electronic patient records. Although FHIR, makes it easier to connect disparate systems, the development community still lacks the knowledge to use it to build browser applications. 
To increase developers’ familiarity with the FHIR standard, a catalogue of JavaScript code snippets that can be used to build SMART on FHIR applications was created. It provides a collation of common features found in clinical/patient apps, such as patient search, medications and observations.

UCL Computer Science BSc Project 2019

Accessibility Technology for the Sight and Sound Hospital

Finding your way around hospitals can be challenging, even with the signage and maps provided. The visually impaired, who often rely on their senses to navigate unfamiliar environments, must orient themselves amidst moving people, background noise, announcements, etc… 
To facilitate indoor navigation for patients with sensory loss, a mobile app was developed that allows users to select a destination using voice commands, and receive turn by turn directions and haptic feedback from NTT DATA Buru-Navi technology for navigation. 

UCL Computer Science BSc Project 2019

HoloRepository: A Repository of Holographic Medical Images

Medical images, such as MRI or CT images, are stacks of 2D images used for analysis and diagnostics. These images are difficult to interpret for non-trained eyes, which could lead to a diminished understanding by patients of their condition and treatment options.
The Holorepository provides an interactive and engaging way to view medical images in 3D. It implements a pipeline to create holographic images from MRI and CT scans. These holograms can then be viewed using devices, such as Microsoft HoloLens, during patient-doctor consultations.

UCL Computer Science MSc Project 2019

My Nephrotic Notebook Mobile App

Parents of children with nephrotic syndrome,or the children themselves use a paper notebook to record daily the level of protein in their urine. These notes are meant to guide patients/parents on medication dosage and inform the clinical team of the patient’s progress, but they are difficult to interpret and often get lost, or damaged. The My Nephrotic Notebook mobile app allows users to easily record proteinuria dipstick readings, and based on these results, it informs users whether they have relapsed, are in maintenance, or have reached a period of remission. The data is stored locally and it can be shared with clinical teams via e-mail.

UCL Computer Science MSc Project 2019

Digital Informed Consent Mobile App

Clinical trials and research studies are vital for treatment innovations. In order to give informed consent, participants are provided with information sheets that describe the study in detail. The current system is paper based and documents must be retained and archived.
The Digital Informed Consent app digitalises this process. It allows for information sheets to be sent via e-mail to potential participants, and for signatures to be taken digitally. All of the digitally completed forms are stored in a database for easy access.

UCL Computer Science MSc Project 2019

Rare Disease Resources

During medical training exposure to patients presenting with rare diseases may be uncommon. The Rare Diseases Resources facilitates the development of a database of rare diseases. It allows medical professionals to upload anonymised medical cases for educational purposes. Trainees are able to view and comment on these cases, thus improving their knowledge and training experience, better preparing them to identify and recognise rare diseases in their future careers.

UCL Computer Science MSc Project 2019

Main

Level 1 40 Bernard Street

WC1N 1LE

London

© 2020 DRIVE