Fingerprint Indexing based on Ridge Orientation and Frequency on GPU

摘要

Fingerprint is a unique characteristic owned by human and is often used as identity. Identifying an unknown fingerprint is done by doing fingerprint matching to a fingerprint database. However, fingerprint matching on a large database needs a lot of time. Therefore, fingerprint indexing is often used to speed up fingerprint identification. Fingerprint indexing consists of creation of index and searching. In this paper, we describe a GPU implementation of the searching phase of fingerprint indexing based on ridge orientation and frequency. The GPU implementation reaches a speedup to 250 times compared to CPU sequential implementation.