Latent fingerprint enhancement and matching using intuitionistic type-2 fuzzy

Abstract

Latent fingerprints are obtained from crime places by law enforcement and forensic agencies to identify the suspect. The latent fingerprints have vague ridge structures and various overlapping valley-based structures that result in low image quality. Furthermore, background noise, low contrast, and low information content make feature extraction difficult. To address these challenges, we propose a novel intuitionistic type-2 fuzzy relation for enhancing image quality. And to improve the matching score, a crossing number method is applied to extract minutiae points. The matching score is calculated using four different distance algorithms: Manhattan distance, Euclidean distance, Earth mover's distance, and chi-square distance. The proposed methodology is evaluated using three datasets IIITD, FVC-2004-1, and FVC-2004-2. The experimental results show that the proposed method yields satisfactory results, with the chi-square distance achieving the highest accuracy. The proposed method can be used by law enforcement and forensic agencies to reduce the workload.