Citation

Askarin, Mohammad Mogharen and Wong, KokSheik and Phan, Raphael (2017) Reduced contact lifting of latent fingerprint. In: 2017 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC), 12-15 Dec. 2017, Kuala Lumpur.

Text 68.pdf - Published Version Restricted to Repository staff only Download (9MB)

Abstract

Fingerprint is the most well-known and successfully deployed biometric modality due to its ease of acquisition, established use, acceptance and high recognition rate (i.e., robustness). One form of fingerprint is called latent fingerprint. Despite its subtle appearance, latent fingerprint is commonly left all over the place unintentionally, including water tap, door knob, elevator button, and cup. To lift these latent fingerprints, the conventional approach involving the process of powdering and taping may physically damage the latent fingerprint. Therefore, a reduced contact method is desirable. This study focuses on latent fingerprints left on curved surfaces, such as water tap, door knob, and water flasks. The latent fingerprint is uncovered (i.e., made visible) by means of fuming, and the end product is captured by a camera. A geometrical compensation method, which takes the curvature of the surface as input, is formulated to geometrically correct (i.e., flatten) the image. The corrected image is further enhanced and sent for matching purpose. Experiments show that the application of the proposed geometrical compensation method is able to flatten the fingerprint image uncovered from a single directional curved surface and improve its matching score

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	Biometric identification
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7800-8360 Electronics
Divisions:	Faculty of Engineering (FOE)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	28 Apr 2021 13:27
Last Modified:	28 Apr 2021 13:27
URII:	http://shdl.mmu.edu.my/id/eprint/7657

Downloads

Edit (login required)