Classifier Combination with Quality Measures
Event details
| Date | 24.03.2010 |
| Hour | 09:15 |
| Speaker | Norman Poh |
| Location | |
| Category | Conferences - Seminars |
Sensory information acquired by pattern recognition systems is invariably subject to environmental and sensing conditions, which may change over time. This may have a significant negative impact on the performance of pattern recognition algorithms.
In the past, these problems have been tackled by building in invarianceto the various changes, by adaptation and by multiple expert systems. Morerecently, the possibility of enhancing the pattern classification system robustnessby using auxiliary information has been explored. In particular, by measuring the extent of degradation, the resulting sensory data quality information can be used with advantage to combat the effect of the degradation phenomena. This can be achieved by using the auxiliary quality information as features in the fusion stage of a multiple classifier system which uses the discriminant function values from the first stage as inputs. Data quality can be measured directly from the sensory data.
Different architectures have been suggested for decision making using quality information. Examples of these architectures are presented and their relative merits discussed. The problems and benefits associated with the use of auxiliary information in sensory data analysis are illustrated on the problem of personal identity verification in biometrics.
Norman Poh is a research fellow at CVSSP, University of Surrey, since 2006, after having completed the Ph.D. degree in computer science conferred by the Swiss Federal Institute of Technology in Lausanne (EPFL). He is one of the work-package leaders in the EU-funded Mobile Biometry (MOBIO) project (led by IDIAP), responsible for designing adaptive multimodal biometric systems. His areas of interest are pattern recognition, video processing, biometric authentication, and information fusion, and he has authored in excess of 50 peer-reviewed publications. He was the recipient of three best paper awards (AVBPA'05, ICB'09 and Pattern Recogition Journal, 2006) and two personal research grants from the Swiss National Science Foundation.
Practical information
- General public
- Free
Contact
- Andrzej Drygajlo