Dr Norman Poh
Qualifications: PhD (EPFL, 2006), MIEEE, IEEE CBP
Phone: Work: 01483 68 6136
Room no: 03 BB 02
Dr Norman Poh joined the Department of Computer Science in August 2012 as a Lecturer.
He received the Ph.D. degree in computer science in 2006 from the Swiss Federal Institute of Technology Lausanne (EPFL), Switzerland. Prior to the current appointment, he was a Research Fellow with the Centre for Vision, Speech, and Signal Processing (CVSSP) and a research assistant at IDIAP research institute. His research objective is to advance pattern recognition techniques with applications to biometrics and healthcare informatics. In these two areas, he has published more than 70 publications, which also include five award-winning papers (AVBPA’05, ICB’09, HSI 2010, ICPR 2010 and Pattern Recognition Journal 2006). Other of his significant achievements include two personal research grants from the Swiss National Science Foundation (Young Prospective Researcher Fellowship and Advanced Researcher Fellowships), and Researcher of the Year 2011 Award, University of Surrey.
I work on pattern recognition with applications to biometrics and healthcare informatics.
If you are interested in any of the topics discussed below, please get in touch with me.
My research on biometrics has been focusing on improving various aspects of a biometric system, which can be regarded as a pattern recognition problem. It consists of the following modules/tasks: a pre-processing module, feature representation, classifier design, and information fusion. Some of these aspects are shown in the figure below.
- Classifier fusion: Combining information from several biometric systems of the same or different modalities can often improve the system performance. Two factors can affect the system performance: correlation and the strength of each classifier. Correlation can be significantly higher in intramodal fusion (combining two classifiers processing the same modality) than in multimodal fusion. So, my prior work consists of advancing the understanding of fusion, taking both aspects into account. This investigation leads to the development of a classifier fusion predictive model giving Equal Error Rate (EER) as output. It is a commonly used metric for biometric authentication because it accounts for the imbalanced nature of genuine and impostor matching.
- Biometric sample quality: Biometric samples collected using a different sensor device or under uncontrolled or adverse environments can compromise the recognition performance of a biometric system. I have developed methods to exploit the signal quality (in the form of "quality measures") to reduce the impact of performance degradation. These methods attempt to adapt the model and/or to calibrate the matching scores.
- Client or user adaptation: Being a human-centric application, researchers working on biometrics have observed the impact of users on the biometric performance. By changing the underlying demographics of the database but keeping the database size (the number of subject) the same, one often obtains a slightly different performance every time. I attempted to understand this phenomenon in two research directions: 1) developing metrics to assess the impact of demographics on the performance; and 2) developing methods to reduce the impact. My second investigation leads me to develop algorithms that adapt to a specific user and later to a group of users sharing some similar characteristics. A client-adaptive system can automatically adjust the appropriate decision threshold based on some training data. Interestingly, directing adjustments for the threshold is more difficult than trying to find a projection function that calibrates the matching score. According to hundreds of experiments that I have conducted, up to a reduction of 50% of EER can be observed.
- Databases and benchmarking: In pursuit of research excellence -- ensuring unbiased experimental reporting and repeatability of experiments -- I have produced and published a number of databases and contributed to benchmarking efforts. Check out and download XM2VTS, BANCA, and Biosecure databases. I have also organised two competitions on biometrics: a benchmark of fusion algorithms in Biosecure workshop 2007 and a face video competition in conjunction with ICB2009.
My future research in this area includes:
- Spoof-robust biometrics: There have been numerous reports about researchers and hackers who were successful at fouling a biometric system by introducing fake biometric samples made from common materials (e.g., using "gummy" fingerprints). There is a urgent need in addressing this security issue. My research will focus on devising algorithms capable of combining liveness detection information into an existing biometric system.
- Self-calibrating biometrics: Biometric systems change over time and its performance is highly dependent on the acquisition conditions. A biometric system that is self-calibrating will be able to adjust its parameters so that it will always perform optimally under all possible operating conditions. My research will focus on tackling this challenge by combining various aspects of adaptation that I have developed into a single framework; these include, quality-adaptation, temporal adaptation, cross-device adaptation, and user-adaptation.
- Exploitation of cohort information: There has been a number of algorithms that rely on a "background" database, or a database of cohort users. For instance, the state-of-the-art approaches to face recognition (e.g., sparse representation and one-shot model) and speaker recognition (GMM) rely on a set of background or cohort users. What is the impact of cohort users when designing a biometric system? How to choose an optimal set of cohort users? Is there an optimal way of incorporating cohort users? These are some of the questions left unanswered in the literature.
There are a number of potential research topics in healthcare informatics. These research topics are shown in the figure below.
- Patient ID encryption: Health records of the same patient are often required to be joined together to form a larger record. They can come from the hospital, primary care clinics (general practices), and disease-specific registries. In order to work with patient data without knowing their identity so as to protect their privacy, we encrypt any data that can reveal their identity, such as, name, address, and any references that that infer their identity (i.e., usually NHS number in the UK). We need to ensure that two encrypted data ("keys") should match each other if they come from the same patient. The challenge here is to produce a unique key even when the identity-related data may have some variation. For instance, even if a patient's name may be spelled slightly differently, we require that the algorithm is still able to identify the resultant keys with high level of probability.
- Ontology: Significant advancement has been made on medical ontology. As a result, different health record systems can communicate with each other thanks to a common set of clinical terms. Five-byte read codes and SNOMED-CT are two prominent examples. Within SNOMED-CT it is now possible to establish relationship among concepts using "is-a" and other entity relationship specifiers. However, this is a tedious and knowledge-driven process. I would like to investigate data-driven methods that are able to improve upon the existing knowledge-driven one.
- Database management and data representation: In order to handle and manage millions of patients of records (which is the case for the Quality Improvement Chronic Kidney Disease -- the QICKD study), we need to encode data in three dimensions: the concept dimensions (in the order of 100K), the patients (millions), and the temporal dimension (up to 20 years of data). Some possible research questions are: 1) How to represent the data in an efficient way? 2) How to present them to clinicians?
- Knowledge discovery and machine learning: Health records are extremely sparse in features and in time. Although there are algorithms capable of dealing with time, such as HMM, the signals are often regularly sampled (e.g., you get 8000 samples every second). In health records, data are not sampled regularly as patients go to their clinicians as and when required or necessary. Conventional temporal models such as HMM is not suitable for this problem; some modifications are required. Furthermore, the method has to be augmented to deal with hundreds of thousands of features that are sparse (where only a few features have data for a given patient). These are some examples of technical challenges to be solved in this research direction..
- Human Computer Interaction(HCI): We need to present meaningful data to clinicians and patients. Both groups of users have very different requirements and purposes. How to reduce information overload when presenting data? This is an art of science in data engineering. Then, we also need to provide a way for clinicians to manipulate the data when carrying a typical seven-minute consultation.
Below are some research topics in healthcare as well as data sets and possible solutions that I have identified. If you spot any topics that you might be interested in, please get in touch with me.
CKD stands for Chronic Kidney Disease; HES: Hospital Episode Statistics -- health data obtained from hospitals in the UK
My research in biometrics is made possible thanks to the following funding agencies:
- FP6 and FP7 EU projects:Biosecure, MOBIO, and BEAT
- Swiss NSF foundation
- Universiti Sains Malaysia for its RLKA fellowship
- Dr. Anil Alexander (Oxford Wave Research Ltd, UK)
- Dr. Samy Bengio (Google Inc., USA)
- Dr. Thirimachos Bourlai (West Virginia University, USA)
- Prof. Josef Kittler (University of Surrey, UK)
- Dr. Krzysztof Kryszczuk (IBM & Patternlab, Switzerland)
- Dr. Sébastian Marcel (IDIAP, Switzerland)
- Dr. Daigo Muramatsu (Seikei University, Japan)
- Dr. Ajita Rattani (Universita di Cagliari, Italy)
- Prof. Fabio Roli (Universita di Cagliari, Italy)
- Prof. Arun Ross (West Virginia University, USA)
- Prof. Massimo Tistarelli (Universita di Sassari, Italy)
- Dr. David Windridge (University of Surrey, UK)
- Members of CVSSP (University of Surrey, UK)
In healthcare informatics, I work with the Clinical Informatics and Health Outcome Research Group.
- R. Wong, N. Poh, J. Kittler and D. Frohlich, Towards Inclusive Design in Mobile Biometry, in the 3rd Int’l Conf. on Human System Interaction (HSI), pg 267-274, 2010 (Best Paper Award in the Human Centered Design track). [pdf]
- N. Poh, R. Wong, J. Kittler and F. Roli, Challenges and Research Directions for Adaptive Biometric Recognition Systems, ICB2009, pg 753-764, 2009. (Best Paper Award) [pdf]
- N. Poh and S. Bengio, Database, Protocol and Tools for Evaluating Score-Level Fusion Algorithms in Biometric Authentication, in Pattern Recognition Journal, Vol. 39, Issue 2, pages 223-233, Feb 2006 (Honourable mention in the Pattern Recognition Journal Best Paper Award 2006). [ps.gz][pdf]
- N. Poh, J. Kittler, S. Marcel, D. Matrouf and J-F. Bonastre,Model and Score Adaptation for Biometric Systems: Coping With Device Interoperability and Changing Acquisition Conditions, Int’l Conf. on Pattern Recognition (ICPR'10), 1229-1232, 2010.(Best Scientific Paper Award in Biometrics and the Human Computer Interaction track) [pdf]
- N. Poh and S. Bengio, A Novel Approach to Combining Client-Dependent and Confidence Information in Multimodal Biometric, in the 5th Int'l. Conf. Audio- and Video-Based Biometric Person Authentication (AVBPA), LNCS 3546, pages 1120-1129, 2005 (Best Student Poster Award). [ps.gz] [pdf]
- P. A. Tresadern, C. McCool, N. Poh, P. Matejka, A. Hadid, C. Levy, T. F. Cootes and S. Marcel, Mobile Biometrics (MoBio): Joint Face and VoiceVerification for a Mobile Platform, IEEE Pervasive Computing, accepted, 2012. [pdf]
- A. Merati, N. Poh and J. Kittler, User-Specific Cohort Selection and Score Normalization for Biometric Systems, IEEE Trans. on Information Forensics and Security, vol 7(4), pg 1270-1277, 2012. [pdf]
- N. Poh and J. Kittler, A Unified Framework for Biometric Expert Fusion Incorporating Quality Measures, IEEE Trans. on Pattern Analysis and Machine Intelligence, 34(1):3-18, 2012. [pdf]
- N. Poh and S de Lusignan, Data-modelling and Visualisation in Chronic Kidney Disease (CKD): A Step Towards Personalised Medicine, Informatics in Primary Care, vol. 19, no. 2, pp. 57-63, 2011.
- N. Poh, J Kittler, C. H. Chan, S. Marcel, C. Mc Cool, E.A.Rua, L. A. Castro, M. Villegas, R. Paredes, V. Struc, N. Pavesic, A.A. Salah, H. Fang, and N. Costen, An evaluation of video-to-video face verification, IEEE Trans. on Information Forensics and Security, 5(4), 781-801, 2010. [pdf]
- N. Poh, D. Windridge, V. Mottl, A. Tatarchuk, and A. Eliseyev. Addressing Missing Values in Kernel-based Multimodal Biometric Fusion Using Neutral Point Substitution. IEEE Trans. on Information Forensics and Security, 5(3):461–469, 2010. [pdf]
- N. Poh, T. Bourlai, and J. Kittler, A Multimodal Biometric Test Bed for Quality-dependent, Cost-sensitive and Client-specific Score-level Fusion Algorithms, in Pattern Recognition Journal, vol. 43, no. 3, pp. 1094–1105, 2009. [pdf]
- N. Poh, T. Bourlai, and J. Kittler, Quality-based Score Normalisation with Device QualitativeInformation for Multimodal Biometric Fusion, IEEE Trans. on Systems, Man, Cybernatics Part B : Systems and Humans, 40(539-554), 2010. [pdf]
- N. Poh, T. Bourlai, J. Kittler, L. Allano, F. Alonso-Fernandez, O. Ambekar, J. Baker, B. Dorizzi, O. Fatukasi, J. Fierrez, H. Ganster, J. Ortega-Garcia, D. Maurer, A. A Salah, T. Scheidat, and C. Vielhauer, Benchmarking Quality-dependent and Cost-sensitive Multimodal Biometric Fusion Algorithms, IEEE Trans. on Information Forensics and Security, 4(4), pp. 849–866, 2009. ;[pdf]
- N. Poh and J. Kittler, Incorporating Variation of Model-specific Score Distribution in Speaker Verification Systems, IEEE Trans. Audio, Speech and Language Processing, 16(3):594-606, 2008. [pdf]
- N. Poh and S. Bengio,Performance Generalization in Biometric Authentication Using Joint User-specific and Sample Bootstraps, IEEE Trans. on Pattern Analysis and Machine Intelligence, 29(3):492-498, March 2007. [ps.gz][pdf]
- N. Poh and S. Bengio,How Do Correlation and Variance of Base Classifiers Affect Fusion in Biometric Authentication Tasks?, in IEEE Trans. on Signal Processing, Vol. 53(11), pages 4384-4396, Nov 2005. [ps.gz][pdf]
- N. Poh and S. Bengio,Database, Protocol and Tools for Evaluating Score-Level Fusion Algorithms in Biometric Authenticationin Pattern Recognition Journal, Vol. 39, Issue 2, pages 223-233, Feb 2006 (Honourable mention in the Pattern Recognition Journal Best Paper Award 2006). [ps.gz][pdf]
I maintain my own publication list (with pdf files) but nowadays automatically indexed publication services are becoming very powerful. I found the following
I deliver lectures, together with Dr Shujun Li, on Web Publishing and Databases (COM1025) to undergraduate students.
With Dr Sotiris Moschoyiannis, we teach the Database and Knowledge Discovery module (COMM033) for post-graduate students.
I also contribute to the Introduction to Multimedia Security module (COMM023) for post-graduate students.
I serve as Postgraduate and Undergraduate coursework coordinator in the Department of Computer Science.
Dr Poh is recognised as an IEEE Certified Biometrics Professional (IEEE CBP). He is a member of IEEE and IAPR.
Working with me
My research focuses on applying pattern recognition techniques to biometrics (e.g., face, fingerprint, iris, speech, and other modalities) and healthcare informatics (processing electronic health records) as discussed in the Research Topics page. If you wish to work with me, consider the following options, depending on your career stage:
The following applies to MSc students taking Computer Science courses offered by the Department of Computer Science, University of Surrey. If you are in this category and are looking for an MSc project, I can offer a few projects.If you are not an existing MSc student at U. of Surrey but consider taking one, check out the postgraduate programme in Computer Science.
For PhD programme, potential candidates are encouraged to apply to appropriate scholarships in their home countries. For research in healthcare informatics, check out NIHR Doctoral Research Fellowship. Before you apply, you must be able to show that you have excellent research and academic track record. Check also the desperate guide for international funding.
For overseas students, please be aware of the following:
- You are likely to pay overseas tuition fees. Some fellowships such as NIHR Doctoral Research Fellowship mentioned above, pay only the Home/EU tuition fee portion. This means that you have to find a way to cover the fee difference. Full-funding for overseas students is rare unless they are truly outstanding.
- You have to satisfy minimum English requirements by taking an English exam.
Applications should come through the online application system. Check out postgraduate research in computer science.
For Post-docs, there are several prestigious fellowships that open annually. The following fellowship schemes are applicable for both research in biometrics and healthcare informatics: Newton International Fellowship, Leverhulme Research Fellowship, Marie Curie Fellowship, EPSRC Postdoctoral Fellowship, and Royal Academy of Engineering Research Fellowship. For research in healthcare informatics, consider NIHR Fellowship and Wellcome Trust Fellowship.
If you want to gain some experience with your sabbatical leave, and would like to work on biometrics or healthcare informatics, please get in touch with me.
My website in EE which is frequently updated.