A Novel Overlapping Fingerprint Partition and Reconstruction Based on Vector Flow Method

Fingerprint is one of human biological characteristic which its uniqueness is proved. In recent year, fingerprint recognition has been becoming an adventure technique, and it is applied on identification for security and privacy regularly. However, it has a problem when fingerprints pattern on the same position. The problem happens from crime scene investigation normally, and it is difficult to recognize. Therefore, how to repair or reconstruct overlapping fingerprints is a problem to research. In this paper, we propose a method to separate and reconstruct two overlapping fingerprints. The Bandelet image inpainting method is combined with contourlet which enhances image and detects gradient vector flow (GVF). And the overlapping part is detected, separated and reconstructed to support criminal identification. In experimental result, the overlapping part of two fingerprints is reconstructed through its non-overlapping part effectively. Our proposed method provides automatic overlapping part detection and reconstruction on two fingerprints processing. It is useful for crime scene investigation.