Dr Imran Mehmood is a SyMeCo postdoctoral fellow with Lero@UoG and is undertaking his fellowship under the supervision of Dr Brian Deegan.

Imran specialises in image quality for camera ISP pipelines, high dynamic range (HDR) imaging, perceptual colour science, and deep learning for autonomous systems. He holds a PhD in Optical Engineering from Zhejiang University, China, where he developed perceptual tone-mapping and natural colour reproduction algorithms for HDR imaging. Before joining SyMeCo, he worked as an AI Imaging Algorithm Engineer at Shenzhen Mikewei Culture Co., Ltd., designing image enhancement pipelines and developing large-scale machine-learning datasets for commercial vision applications. He has published peer-reviewed works in leading journals and conferences, including IEEE Access, Optics Express, JOSA A, and the Colour and Imaging Conference, covering HDR datasets, perceptual tone-mapping, colour correction, and image-quality evaluation methods. His research focuses on camera ISP pipelines, image quality, deep learning and computer-vision perception for autonomous driving under complex illumination conditions.

Imran’s SyMeCo research project, titled “Dynamic Contrast Management and Colour Constancy in Autonomous Vehicles for Real-Time Applications” will develop a comprehensive HDR and raw image dataset that captures real world driving conditions, along with deep learning models optimised for contrast enhancement and color constancy. The research incorporates human visual system modeling into computational imaging to improve the robustness and reliability of perception in autonomous driving systems.

Expected impact – This research is expected to significantly improve visual perception in autonomous vehicles by ensuring reliable contrast and colour accuracy under all ambient lighting conditions. The dataset and models developed will: (i) Enhance object/lane/sign recognition in challenging illumination, (ii) Reduce perception failures that can compromise safety,
(iii) Provide the community with the first HDR-focused autonomous driving image resource, and (iv) Accelerate innovation in both academic research and industrial implementation

Ultimately, the impact aligns with safer and more trustworthy autonomous mobility in diverse real-world environments.

Interdisciplinary aspects – The project is inherently interdisciplinary, combining:

• Computer Vision and Deep Learning for development of tone mapping algorithms and autonomous perception models
• Colour and Optical Science for implementation of perceptual models of human vision for accurate colour reproduction
• Autonomous Vehicle Systems Engineering for real time integration testing and evaluation of imaging algorithms
• Data Science for design curation and management of structured HDR datasets

Collaboration with industrial partners such as Valeo and Future Mobility Campus Ireland further reinforces its multi-sector application focus.