The objective of this research project is to search for new algorithms for low quality fingerprints segmentation, including latents. Latents images are extracted from uncontrolled environments, usually obtained in places that have been touched by suspects in a crime scene. The goal of segmentation is to make the process of analysis of fingerprint images as automatic as possible, increasing the performance of systems that make fingerprint recognition among the target with corresponding fingerprints registered on a database of suspected citizens. The segmentation of fingerprints obtained in controlled environments is a problem practically solved, however the segmentation of low quality fingerprint is still an open problem.
News published in Agência FAPESP Newsletter about the scholarship: