MSc Diploma Project
Improve the speed and robustness of fingerprint technology
Céline Di Venuto
Prof. T. Ebrahimi
The goal of the project is to explore different matching approaches in order to further improve the detection speed of the technology. In particular, hierarchical approach will be examined as well as fast image matching technique. One particular technique is a pyramidal matching. The goal is to match down sampled versions of the images to classify them in a set of “possible matches”. The matching for higher size images is done only for images in the set of “possible matches”. Finally, the algorithm must be able to cancel the various noises of the camera (spatial and temporal) as well as the impefections of the imaging system (focus, positonning, etc).
All these techniques will be implemented in an application that can be installed on a customer PC connected to a microscope. When the user takes the picture with the microscope, the application will automatically detect if the object is or not registered. The application will use multiple threads to be able to take advantage of the performance of multi core machines.
False positives and false negatives will be tested thanks to an automatically generated database of images. The images should be random but resembling to the real images.
It has also been remarked that the camera of the microscope induces some noise which should be removed to improve the correlation of images.