GEMTELLIGENCE: Accelerating gemstone classification with deep learning.

ABSTRACT: The value of luxury goods, particularly investment-grade gemstones, is influenced by their origin and authenticity, often resulting in differences worth millions of dollars. Traditional methods for determining gemstone origin and detecting treatments involve subjective visual inspections and a range of advanced analytical techniques. However, these approaches can be time-consuming, prone to inconsistencies, and lack automation. Here, we propose GEMTELLIGENCE, a novel deep learning approach enabling streamlined and consistent origin determination of gemstone origin and detection of treatments. GEMTELLIGENCE leverages convolutional and attention-based neural networks that combine the multi-modal heterogeneous data collected from multiple instruments. The algorithm attains predictive performa

SUBMITTER: Bendinelli T 

PROVIDER: S-EPMC11336078 | biostudies-literature | 2024 Aug

REPOSITORIES: biostudies-literature