Dr James Howard

I achieved my second Kaggle gold medal (and my first solo) in the SETI Breakthrough Listen challenge. Fascinating competition using AI to process radio telescope signals. Mini write-up here https://kaggle.com/c/seti-breakthrough-listen/discussion/266460

Interested in how these "deep learning" systems appearing in your echo/CMR/CT software actually work? Our review in @Heart_BMJ explains what's going on "under the hood" in neural networks, with a focus on medical imaging: https://heart.bmj.com/content/early/2021/07/23/heartjnl-2020-318686

Our AI collaborative's first paper is out in @circimaging! We trained an AI to perform measurements of parasternal long axis #echofirst images, using expert opinions from cardiologists and physiologists from 17 hospitals across the UK. But hasn't this been done before? Well, yes... but not…

We investigated whether deep learning of the sequences acquired in the first minutes of a scan could provide an early alert of abnormal features. It does this by segmenting out slices of the heart acquired at the start of a scan, and using these to…

Thank you everyone who made SAMSON happen, but by far most of all Frances Wood. It's truly her trial. We are so grateful from the @TheBHF and @ImperialNHLI's support. You can read our simultaneously published research correspondence in NEJM here https://nejm.org/doi/full/10.1056/NEJMc2031173?query=main_nav_lg…

Ecstatic to say our 2-man team ended up finishing 5th in Kaggle's and Facebook's DeepFake Detection Challenge! It was my first AI competition, and the $40,000 prize was a wonderful surprise!

The identification of adverse pressure signals during an angiogram, termed "damping", is necessary to maintain safety and the accuracy of coronary physiology measurements. We developed a convolutional neural network which is able to accurately monitor these signals. This work has been published in JACC: Cardiovascular Interventions

We've developed a free-to-use tool utilising a convolutional neural network which is able to accurately identify the model of pacemaker present on a chest X-ray. This work has been published in JACC: Clinical Electrophysiology (open access).

Automated echocardiographic interpretation hinges on the correct recognition of the view (imaging plane and orientation). Current state-of-the-art methods for identifying the view computationally involve 2-dimensional convolutional neural networks (CNNs) and ignore information describing the movement of structures throughout the cardiac cycle. Here we explore…