This research paper addresses the task of image-based identification of precious stones, focusing on the integration of edge computing devices, specifically the Raspberry Pi 5, and utilizing the YOLOv8 instance segmentation model for gemstone identification. The study explores the challenges associated with classifying diverse gemstones and leverages the computational capabilities of the Raspberry Pi 5 for efficient edge processing. We present a detailed investigation into the design and implementation of the YOLOv8 instance segmentation model on the Raspberry Pi 5, highlighting the model’s efficacy in accurately identifying and classifying various precious stones based on image input. The experimental results demonstrate the feasibility of deploying sophisticated image-based classification systems on edge devices, with a particular emphasis on gemstone recognition. The combination of edge computing and advanced instance segmentation techniques contributes to the development of a portable and efficient solution for gemstone identification, presenting opportunities for real-world applications in industries such as jewelry authentication and gemology.