When suspects of crimes are caught by the police, their facial images and fingerprints are recorded and added to the extensive databases they have. If on a crime scene fingermarks are found they are compared to the databases using advanced fingerprint matching algorithms and in this way it is possible to obtain a short list of suspects quickly. However, when the fingerprints are recorded at the police office, sometimes traces of other fingerprints are left on the glass plate of the fingerprint sensor. Also, sometimes, the calibration of the sensor is not performed correctly and a "negative" of the previous fingerprint is stored. The left traces are mixed with the new fingerprint forming a ghost fingerprint pattern, often called "ghost" for short. In the databases of the police such ghosts are present, but it is not known how many and how serious the effect is. Of course they interfere with the proper recognition of fingerprints, sometimes giving a wrong match, sometimes causing a failure to match a fingerprint. This presentation gives an overview of the various approaches we attempted to detect these ghost fingerprints and estimate their impact on the matching process.
Dr. Luuk Spreeuwers studied Electrical Engineering at the University of Twente, Netherlands. In 1992 he obtained his PhD from the University of Twente. The title of his PhD-thesis is: Image Filtering with Neural Networks: Applications and Performance Evaluation. Subsequently Luuk Spreeuwers worked at the International Institute for Aerospace and Earth Sciences (ITC) in Enschede, Netherlands, the University of Twente in a SION project on 3-D image analysis of aerial image sequences and in Budapest at the Hungarian Academy of Sciences in a 3-D textures ERCIM project. From 1999-2005 Luuk Spreeuwers worked on 3-D modeling and segmentation of the human heart in MRI at the Image Sciences Institute of the University Medical Centre in Utrecht, the Netherlands. Currently he is an Associate Professor at the Data Management and Biometrics (DMB) group at the University of Twente and is involved in projects on 2D/3D face recognition, face morphing, finger vein biometrics and others. Luuk Spreeuwers has published over 80 papers at international conferences and in journals. His expertise involves digital image processing and analysis, medical image analysis, biometrics and pattern recognition in general.