Dr Peter Yuen

Areas of expertise

• Bioinformatics
• Biomedical Engineering
• Biosensors & Diagnostics
• Computing, Simulation & Modelling
• Counter-IED
• Defence Sensors
• Environment and Health
• Food Safety
• Industrial Automation
• Instrumentation, Sensors and Measurement Science
• Monitoring and Environmental Informatics
• Nanotechnology
• Sensor Technologies
• Soil

Background

Peter Yuen obtained a First Class Honours in Physics and subsequently a PhD in Semiconductor Physics from Imperial College, University of London, in 1987. Peter was accredited Fellow of the Institute of Physics (FInstP) and the Institute of Mathematics and its Applications (FIMA) in 2001. Between 1997 to 2007 he worked with corporate laboratories such as BAE Systems, Defence Evaluation and Research Agency (DERA) and Asahi Kasei (Japan). Since 2013, Peter has been a Reader in remote sensing & electro-optics at Cranfield University. 

Peter has over 90 peer-reviewed journal and conference papers on topics including semiconductor physics, sensors, defence science, electro-optics, image processing including remote sensing, hyperspectral imaging, machine vision and machine learning disciplines, with over 530 journal citations from his recent 5 year publications since 2018 (H-index 11). Peter contributed 5 of his best journal papers towards the REF2020 entry and he is an active peer reviewer for high impact factor journals such as the IEEE Transactions on Geoscience and Remote Sensing (TGRS), Neural Computing and Applications (NCAA), Optics & Laser Technology (JOLT), Optical Engineering (OE), MDPI’s Remote Sensing, Computers and Electronics in Agriculture (COMPAG) etc. 

Peter’s current research includes computer vision using statistical machine learning techniques, as well as artificial intelligence deep learning methods for the detection of 'difficult' targets from heavily cluttered and camouflaged backgrounds.

Research opportunities

Peter’s team has strong collaborations with universities in the UK and world-wide, for the R&D of remote sensing technology specifically for agricultural, environmental, industrial and security applications. Collaborators are mostly academic institutes in the UK, geared for the development of next generation sensor and sensing system for surveillance and world-monitoring applications towards the clean air objective. The team has strong links with a few agricultural universities in the UK (e.g., Aberystwyth University, Rothamsted Research, etc.) and in India/Pakistan. Collaborations have taken place through joint projects funded by government bodies such as EPSRC, NERC, Royal Society, DSTL and Newton Fund, etc. We welcome full time, part time and exchange researchers and students to work with us, either on-site or remotely.       

Current activities

Current research topics in the team include:

• Electro-optics research for the development of efficient snap-shot multispectral imaging system.
• Compressive imaging algorithm development to achieve high degree of spectral reconstruction accuracy from the snap-shot multispectral imaging data.
• Hyperspectral scene simulations and their validations: to acquire an end-to-end realistic hyperspectral scene simulation capability within the next 5 years.
• Reduction of dimensionality in hyperspectral imaging through band selections for the enhancement of efficacy and efficiency of detecting ‘difficult’ targets.
• Data driven optimisation of deep learning artificial intelligence methodology for the detection and classification of high dimensional data.
• Efficacy, efficiency, robustness and flexibility of deep learning AI methodology in comparison with statistical machine learning algorithms for target detections in high dimensional hyperspectral imaging data.
• Efficiency of X-ray imaging using coded aperture compressive sensing technique.
• Robustness and efficiency of deep learning for the analysis of big data which consists of multi-modal of imaging data.
• Atmospheric correction of hyperspectral scene in the wide spectrum ranging from visible to near infra-red and far-infrared domains.
• Illumination independent targets detection methodology development.
• Diffuse optical imaging for biomedical applications such as in-vivo blood oxygenation assessment, tumour classification, affective computing.
• Spectral constancy including shadow removal.
• Tomography including the development of x-ray back scatter imaging system.
• Hyperspectral diffused optical tomography for target recognition in undersea environments.
• Colour constancy for improving image visualisation and target detection.
• Automatic target recognitions using cortex-like machine vision in the spectral and spatial domains.

Clients

Publications

• Articles in journals
• Conference papers
• Books