Applied Vision

Explainable AI for Efficient Hyperspectral Band Selection in Textile Recycling: A Score-CAM Approach

Authors: Achraf Guenounou (University of Udine) , Harald Ganster (Joanneum Research, Graz) , Jean-Philippe Andreu (Joanneum Research)

Abstract

To enable the transition to cost-effective, real-time multispectral sensors, this study
introduces a novel Explainable AI (XAI) framework for spectral band selection
by adapting Score-CAM—typically used for 2D images—to 1D hyperspectral
signals. This XAI-driven approach is rigorously evaluated against established
chemometric and machine learning baselines, including Weighted Regression
Coefficients (WRC), Variable Importance in Projection (VIP), and the Successive
Projections Algorithm (SPA). This method reduces data volume by over 90% while
matching full-spectrum baseline performance. Ultimately, this research validates
XAI as an interpretable, robust tool for designing efficient, low-cost optical sorting
systems for the circular economy.

Keywords:

How to Cite: Guenounou, A. , Ganster, H. & Andreu, J. (2026) “Explainable AI for Efficient Hyperspectral Band Selection in Textile Recycling: A Score-CAM Approach”, Proceedings of the Austrian Symposium on AI, Robotics, and Vision. 3(1).