Fully Funded PhD in Vision–Language and Geometry–Language Models for Cardiac CT Report Generation Imperial College London in United Kingdom

Imperial College London’s BioMedIA Team MIRA is advertising a fully funded PhD studentship in the AI for Healthcare Centres (AI4Health), co-funded by industry partner Heartflow.

The project focuses on vision-language and geometry-language models for grounded report generation from cardiac CT. The research aims to fuse 2D images, 3D volumes, and geometric structures such as centerlines and meshes to build clinically grounded, explainable systems that can generate reports and answer questions about coronary CT angiography scans, with each finding linked back to precise anatomy.

Students will work with large-scale open datasets including MIMIC-CXR, PadChest-GR, CT-RATE, and RadGenome-ChestCT, and will gain experience in medical imaging, machine learning, computer vision, deep learning, and Python.

Supervision is by Ben Glocker, Dr Chen Qin, Dr Wenjia Bai, and Dr Matthew Sinclair (Heartflow).

Eligibility highlights: applicants should have a background in CS, ML, or AI; engineering, maths, or physics backgrounds with strong data-analysis skills are also welcome. Prior medical imaging experience is helpful but not required.

Funding: fully funded for Home (UK) students, including tuition and a tax-free stipend of at least £23,805 per year for 3.5–4 years, plus conference travel support. A limited number of overseas studentships are available for international applicants.

Start date: October 2026. Apply via the AI4Health portal and review the programme details before submitting.