Smart optical design: AI tools for the next generation of cinematic camera lenses

This project aims to develop a computational framework combining computer vision, computer graphics, and machine learning to accelerate and improve the design and simulation of camera lenses. Traditional lens design is slow, labour-intensive, and relies on expert-led iterative methods. We propose instead to learn mappings between lens design parameters and image-quality metrics using data-driven, differentiable rendering models. The framework will support two modes: forward simulation (predicting the imaging output of candidate lens configurations) and inverse design (proposing lens architectures to meet a target imaging/perceptual objective).

The industrial partner for this project is Cooke Optics, a UK-based, globally recognised manufacturer of precision cinematographic lenses. Cooke will contribute expertise, real-world data, and validation tests to bridge the gap between academic methods and practical lens systems. In collaboration, we will benchmark the learned models against Cooke’s proprietary lens designs and deploy prototypes within their design workflows. The project’s outcomes include new algorithms, open-source simulation tools, and pathways to technology transfer. This work aligns strongly with EPSRC’s priorities in AI, robotics, and engineering, while promoting UK competitiveness in computational optics.

You will need to meet the minimum entry requirements for our PhD programme.

We are seeking a highly motivated and technically strong candidate with a background in computer science, electronic engineering, applied mathematics, physics, or a closely related discipline. The ideal candidate will have:

• A first-class or high 2:1 honours degree (or equivalent) and preferably a Master’s degree in a relevant field.
• Strong mathematical foundations, particularly in linear algebra, optimisation, probability, and numerical methods.
• Experience in computer vision, computer graphics, computational imaging, or machine learning.
• Excellent programming skills in Python and/or C/C++, with experience using machine learning frameworks such as PyTorch or TensorFlow.
• Familiarity with rendering, simulation, or physics-based modelling (e.g., ray tracing, differentiable rendering, optics, or photometric modelling) is desirable.
• An interest in interdisciplinary research at the interface of AI, computational imaging, and industrial optical design.

We particularly welcome applicants who are excited about translating cutting-edge AI research into real-world engineering systems through close collaboration with industry. Strong problem-solving ability, creativity, and the capacity to work both independently and within a collaborative research team are essential.

Prior experience with optical systems, inverse problems, generative models, or high-performance computing would be advantageous but is not required.

Open to any UK or international candidates. Up to 30% of our UKRI funded studentships can be awarded to candidates paying international rate fees. Find out more about eligibility.