Chemical mastery to AI precision: Advancing forensic fingerprint analysis | Poster Board #569

Forensic chemistry has evolved by the progressive sophistication of methods for uncovering material truths, notably in fingerprint identification. The talk will delve into this advancement, tracing the trajectory from chemical innovations to the integration of artificial intelligence (AI) for enhanced fingerprint detection and data analysis. Fingerprinting was one of the cornerstones of forensic identification, being beneficial from a repertoire of chemical agents (e.g. fuming iodine, I₂) to nanostructured metal oxides (e.g. titanium dioxide, TiO₂) for enhanced visualization under varying substrates.

The resurgence of AI has revolutionized fingerprint analysis through deep learning algorithms. Convolutional Neural Networks (CNNs) can facilitate automation of data evaluation with high throughput of fingerprint samples due to their efficacy in image recognition tasks. These AI models enabled recognition of fingerprint patterns within the intricate whorls and ridges of fingerprints. Furthermore, the partial or smudged prints can be evaluated through Principal Component Analysis (PAC), to separate the prints, as a result, enhancing the accuracy of matches.

This talk will showcase the symbiosis between chemical methodologies and AI in forensic fingerprint analysis, with a focus on precision and efficiency gained by our previous research activities. The previous data showed that the Chitin-shielded mixed iron oxides showed 40% improvement in smudge print detection using CNNs based PCA analyses, compared to graphitic carbon, based on FBI laboratory testing (Fairfax VA). This presentation offer a narrative, highlighting the significant milestones in fingerprint analysis from the perspective of forensic chemistry.