Student research project
Can a visual scene predict circadian light exposure?
From images of an environment to estimates of biologically relevant light.
The question
A photograph describes how an environment looks, but can it help estimate the light that matters for circadian physiology? The challenge is to connect visual information with a physical exposure measure.
The project
Investigate computational approaches that relate visual scenes to reference light measurements. The work will examine what image information supports prediction, how uncertainty should be represented, and where models may fail when conditions change.
What you will do
- Read across computer vision, radiometry and circadian light measurement.
- Build a documented image-analysis or modelling workflow for an agreed dataset.
- Evaluate predictive error and robustness, and interpret the limitations of image-based estimates.
Possible scope
A thesis or engineering project could develop and evaluate a modelling approach. A shorter placement could focus on data quality or a bounded comparison. Experience with quantitative programming and image processing is useful; specific technical implementation is discussed individually.
How to apply
Send a short motivation, CV, relevant skills, preferred project, dates, format and location. Mention external funding. Contact Prof. Manuel Spitschan using the address below.
Location, timing and scope are agreed with the supervisor; availability depends on capacity and the work involved.