Artificial Intelligence-Assisted Electrochemiluminescence Identification of Latent Fingerprints

Latent fingerprints (LFPs) are critical evidence in forensic science, yet their identification usually requires specialized visualization methods that are strongly influenced by the properties of the substrates on which they are deposited. Achieving noninvasive and highly reproducible visualization of LFPs remains a significant challenge. Here, we report the visualization of LFPs on graphite paper (GP) using highly surface-sensitive electrochemiluminescence (ECL) imaging. The imaging mechanism relies on the localized electrochemical inertness of lipid residues on fingerprint ridges, which spatially confines ECL emission to the furrows (uncovered regions), thereby generating a high-contrast negative image with dark ridges against a bright background. In addition, owing to its excellent flexibility and strong lipid-binding capability, GP can serve as both a conductive imaging substrate and a transfer medium for lifting LFP residues from a wide variety of objects, followed by direct ECL visualization without additional treatment, greatly expanding the applicability of this method for personal identification. By further integrating artificial intelligence, we developed an ECL fingerprint identification system that includes database construction, model training, and fingerprint matching, ultimately enabling highly accurate identification of LFPs transferred from 15 types of contact surfaces with matching confidence values greater than 0.94.