Research Assistant/Associate in Image Quality for Machine Learning and Optimization
Department of Computer Science and Technology, West Cambridge
A position is available to work on visual models and metrics for deep learning and optimisation. The project aims to develop a family of image/video cost functions and evaluation metrics that predict the perceived degradation of visual quality while accounting for important factors like display size or colour gamut. We are especially interested in novel display modes, such as AR and VR. This work is done in a collaboration with our industrial partner, Meta Reality Labs.
Developing improved cost functions and metrics could accelerate research on deep learning and optimization-based rendering techniques. It could have a major impact on visual computing and improve algorithmic performance for important cases like AR and VR. We also aim to understand the theoretical and practical aspects of making a given cost function robust and applicable in practical scenarios.