Leveraging JMP Pro and Torch Add-In for Mineral Classification (2025-US-PO-2313)

This poster was inspired by JMP’s growing number of mining customers in Latin America.

Identifying minerals in the field can be challenging, often requiring specialized knowledge yet still being susceptible to mistakes. Traditional methods depend heavily on manual feature extraction and microscopic examination, which can be labor-intensive, time-consuming, and prone to inaccuracies. This poster introduces an innovative approach using deep learning within JMP Pro and the Torch Deep Learning Add-In to classify minerals more effectively. The predictive model developed through this method could be implemented in a mining environment using cameras and other equipment to quickly classify minerals on site.

We import folders containing mineral images into JMP and apply the Torch Deep Learning Add-In to build a predictive model. This approach eliminates the need for manual feature extraction and coding, offering a more streamlined and efficient solution for mineral classification. Our method simplifies the identification process and improves accuracy, making it accessible to a wider audience without requiring specialized expertise.

The poster examines the challenges with traditional methods of mineral identification, emphasizing the drawbacks of conventional methods. It then details the steps taken to import and preprocess the data set, followed by the fitting of the deep learning model. Finally, we provide an analysis of the results, showcasing the model's effectiveness and its potential benefits for the mining sector.