Standardising Spectral Imaging in Minimally Invasive Surgery
Project ID: 2531bd1650
(You will need this ID for your application)
Research Theme: Healthcare Technologies
Research Area(s):
Healthcare technologies theme
Artificial intelligence technologies theme
Artificial intelligence and robotics theme
UCL Lead department: Computer Science
Lead Supervisor: Danail Stoyanov
Project Summary:
Why this matters?
In keyhole (minimally invasive) surgery, tissue can look normal under white light even when blood supply is poor. Hyperspectral imaging (HSI) measures many wavelengths at once and can reveal unique signs of compromised perfusion. However, the measured signal drifts as the camera moves and lighting changes. This project makes HSI dependable in motion by stabilising the measurement itself, so surgeons get steady, interpretable maps during real procedures.
What you’ll do?
- Build a HSI module paired with a light-spectrum sensor and camera depth input.
- Develop deep learning algorithms for overcoming common problems is measured signals, such as spectral cross talk and inaccuracy due to artefacts.
- Develop real-time calibration pipeline that separates lighting from the tissue signal, correcting for the signal drift, vignetting, distance and angle, and handling glare.
- Output real-time reflectance and derived oxygenation maps with clear quality and uncertainty indicators.
- Validate from phantoms to non-interventional live surgery, benchmark latency and agreement with reference measurements.
What you’ll learn
Optics and spectral imaging, 3D computer vision, Deep learning, GPU-accelerated computing, and practical clinical study methods.
Who should apply?
Applicants from engineering, computer science, physics or related fields. Useful experience includes imaging/optics, computer vision and medical robotics with interest in biomedical applications and interdisciplinary research.
Why This Project?
- Work alongside experts in computer science, biomedical engineering, and surgery at the UCL Hawkes Institute. Become part of the EPSRC Optical and Acoustic imaging for Surgical and Interventional Sciences (OASIS) hub.
- Contribute to technology that can improve surgical precision and enhance patient outcomes across multiple surgical specialities.
- Gain expertise in optical imaging, computer vision, deep learning, high-performance computing.
Join us in pushing the boundaries of surgical technology and making a significant difference in healthcare through innovative research.