Description
Project Goals
Join us in building the UK's largest and most diverse open-access biobank of echocardiographic images and annotations. This collaborative effort will fuel the development of AI algorithms that revolutionise cardiac care through improved image processing, acquisition, and interpretation..
Active Learning
The images in the biobank project are selected using an active learning approach, which prioritises the most uncertain or informative cases for annotation. This method ensures that clinicians focus on labeling the images that will most effectively improve the AI model's performance, reducing the overall number of images that need to be annotated. More deatails can be found here: sciencedirect.com
Quality Ranking
For projects focused on image quality assessment, we utilize a Glicko-2 rating system. This method, originally designed for competitive games, allows for precise and dynamic ranking of images based on relative quality comparisons made by expert clinicians. Each image is assigned a rating, and the system adjusts these ratings based on the comparative assessments provided by the clinicians. More deatails can be found here: Medical Artificial Intelligence
Unity Labelling Platform
Our streamlined labelling platform works seamlessly across devices (Safari for MacOS, iPhone, iPad; Google Chrome for Windows, Linux, Android). The platform auto-detects your device type and provides tailored installation instructions. Secure sign-in with a Google account is required. The platform is freely available at: unityimaging.net
Who Should Participate?
- Echocardiographers: Your expertise in image interpretation is crucial for accurate assessment.
- Cardiologists: Your clinical knowledge can help understand the impact of foreshortening, endocardial borders visibility, etc on diagnosis.
- Sonographers: Your experience in image acquisition can provide valuable insights into optimizing image quality.
Your Role
- Review Echocardiographic Videos/Images: You will be presented with a series of echocardiographic videos/images.
- Assessment and Ranking: Each project has a demonstration video explaining how to access the project on unityimaging.net along with the guidelines. Using the guidelines provided and your expertise, you will assess each video/image.
- Provide Feedback (Optional): Share your valuable insights and suggestions for improving the annotation process or the assessment criteria.
Code Repositories
In the wake of open science, our priority has been the full publication of details of all methods developed in research so that they can be reproduced, criticised, and improved upon. We make all our research materials, including the datasets, code of experiments, and analyses, freely available at our GitHub page: github.com/intsav
Datasets
The dataset for each project is described/provided in the COMPLETED PROJECTS section.
Work with us
- Contribute to groundbreaking research: Your expertise will directly impact the development of cutting-edge AI algorithms that can revolutionise cardiac care.
- Earn competitive compensation: Get paid for your valuable time and expertise.
- Flexible work schedule: Work from the comfort of your own home and set your own hours.
- Make a real difference in patient lives: Help improve the accuracy and efficiency of echocardiographic diagnosis, leading to better outcomes for patients worldwide.
- Expand your professional network: Collaborate with leading experts in the field of echocardiography and AI.
- Enhance your skills and knowledge: Gain valuable experience in image annotation and contribute to the advancement of medical imaging technology.
.jpg)
.png)
