A slap fingerprint based verification system invariant to halo and sweat artifacts

The paper proposes an efficient slap fingerprint based verification system. It can separate multiple fingerprints of a subject merged due to halo and sweat by applying model driven artificial intelligence. It makes use of the modeled orientation field to determine whether a segmented image consists of single or multiple fingerprints. Further, it explores the possibility to split the multiple fingerprints using sum of residual errors in the actual and modeled orientation field. Experimental results are analyzed on a database of 1200 slap-images out of which 400 slap-images contain halo and sweat artifacts. An equal error rate of 5.31% has been achieved for slap-image matching. Experimental results show the superiority of the system over other existing well known systems.