Dr Alexandra Psarrou

Dr Alexandra Psarrou


I am a Reader in the School of Computer Science and Engineering having served as Head of School from 2002-2019. I am a member of the Computational Vision and Imaging Technologies Research Group.

I received my BSc in Computer Science (1987) and MSc in Advanced Computer Science (1988) from Queen Mary University of London. Following my graduation I worked as a Knowledge Engineer on an AI assisted system (AICQS) for the support of UNISYS customer services (1988-1990) and as a Research Fellow on an SERC medical image interpretation project for the dynamic modelling of cancerous cells (1990-1992). The latter project initiated my interest in motion-based recognition and the analysis of visual behaviour. I received my PhD from Queen Mary, London in 1996 with a thesis on the development of parallel implemented artificial neural networks on DAP for motion-based recognition.

 At the University of Westminster I founded the Computational Vision Research Group (CVRG) in 1997. CVRG’s early work focused on the use of dynamic visual information in the modelling and interpretation of objects and events in a scene, based on recurrent neural networks, multi-view Kernel PCA models and data-driven models using minimum description length. This position was also argued through the publication of a bestselling monograph “Dynamic Vision: From Images to Face Recognition” (Gong, McKenna, Psarrou 2000).  In later years I worked on hypersprectral analysis for applications in cultural heritage, video annotation and unsupervised region extraction and abstraction. Most recently I have been involved with the development and evaluation of explainable deep learning models for image system performance and the metrification of image quality for automotive applications.

Throughout my career at Westminster I have been involved in the development of teaching and research strategy. During my tenure as Head of School I have successfully delivered change management through two University restructures. I have  University-wide experience in academic processes and have acted as a member of  a number of University and Faculty committees including: the University’s Research Committee (1997-2003), University’s Work Allocation Modelling Committee (2012-2018), University’s Curriculum Review and Innovation Committee (2016 – 2018), member of University’s Athena SWAN  SAT (2012-2018), elected member of Faculty’s Research Committee (2014-2018),  Chair of the Computer Science – Engineering Athena SWAN SAT (2017-2018), member of the School/Campus/Faculty Executive Group (1997-2018).  I serve as member of assessment boards and validation panels both externally and internally within the University of Westminster and I have acted as PhD external and internal examiner.  I am a reviewer for EU projects and Advanced HE panel member for Athena SWAN applications.

Fellowships

Senior Fellow of the Higher Education Academy (SFHEA)

Fellow of the British  Computer Society (FBCS)

Research grants

NOESIS - Non-dEStructive Image- based manuscript analysis System, European Commission, 6th Framework Programme, 1 September 2004 - 31 August 2007, Grant Value: EUR 709,850. Grant value for UoW: EUR 382,000 Role: Project  Coordinator, Principal Investigator and Work Package Leader.

diARTgnosis - Study of European Religious Paintings, Culture 2000, Cultural Heritage. 1 November 2000 - 1 November 2003, Grant Value: EUR 518,782, Role:  Co-Principal Investigator for UoW.

HISTORIA - Heraldic Images Storing Applications, 3rd Framework Programme, 1 February 1995 - 30 April 1996, Grant Value for UoW: £62,000, Role: Co-Principal Investigator for UoW and Work Package Leader.

1998 – 2002 (£90,000) Principal Investigator of HEFCE NFF research collaboration project with Queen Mary, London on Temporal Models of Human Faces and Gestures


My  early work focused on the use of dynamic visual information in the modelling and interpretation of objects and events in a scene, based on recurrent neural networks, multi-view Kernel PCA models and data-driven models using minimum description length. This position was also argued through the publication of a bestselling monograph “Dynamic Vision: From Images to Face Recognition” (Gong, McKenna, Psarrou 2000).  In later years I worked on hypersprectral analysis for applications in cultural heritage, video annotation and unsupervised region extraction and abstraction. 

My current research interest ficus on the the development and evaluation of explainable deep learning models for image system performance and the metrification of image quality for automotive applications.

My main areas of interest are:

  • Deep learning methods for Image Quality Performance
  • Explainable learning models 
  • Reinforcement learning  
  • Hyperspectral analysis and modelling of inks and paintings for cultural heritage  

Research grants

NOESIS - Non-dEStructive Image- based manuscript analysis System, European Commission, 6th Framework Programme, 1 September 2004 - 31 August 2007, Grant Value: EUR 709,850. Grant value for UoW: EUR 382,000 Role: Project  Coordinator, Principal Investigator and Work Package Leader.

diARTgnosis - Study of European Religious Paintings, Culture 2000, Cultural Heritage. 1 November 2000 - 1 November 2003, Grant Value: EUR 518,782, Role:  Co-Principal Investigator for UoW.

HISTORIA - Heraldic Images Storing Applications, 3rd Framework Programme, 1 February 1995 - 30 April 1996, Grant Value for UoW: £62,000, Role: Co-Principal Investigator for UoW and Work Package Leader.

1998 – 2002 (£90,000) Principal Investigator of HEFCE NFF research collaboration project with Queen Mary, London on Temporal Models of Human Faces and Gestures

Awards

1999    Best Scientific Paper Award, British Machine Vision Conference

2008    Best paper award, International MultiConference of Engineers and Computer                Scientists 

2020    Best paper award, IS&T International Symposium on Electronic Imaging:Image              Quality and System Performance XVII : Displaying, Processing, Hardcopy, and              Applications


  • Computational Vision and Imaging Technology