Latent Fingerprint Database Using Reflected Ultra Violet Imaging System

Abstract

Latent fingerprint is considered as the key evidence during crime scene investigations. Various powder and chemical methods are available for visualizing the latent fingerprints as these finger impressions are not directly visible through the naked eye. However, these methods may damage the finger impressions in case if they are not properly lifted and handled carefully. Preserving the evidential value of the located latent fingerprints, hence becomes pivotal for analyzing and identifying the suspected individual. Nowadays, optical touchless technology is being prevalent for developing and visualizing the latent finger impressions. Reflected Ultra Violet Imaging System (RUVIS) is one such optical touchless device. There are number of powder based latent fingerprint databases available. However, database of latent fingerprints using optical touchless technology is not available in the literature. The paper presents the latent fingerprint database developed and captured using the touch-less acquisition device (RUVIS). Extraction of level 3 features from the latent fingerprints plays a significant role in matching these latent impressions with plain impressions. Further in this paper, level 3 features, particularly pores are extracted using Fully Convolution Neural Network (FCN) from the collected latent fingerprints using the RUVIS.