An explainable segmentation decision tree model for enhanced decision support in roundwood sorting

Julia Achatz, Pauline Sailer, Sven Mayer, Mark Schubert

Abstract: Highlights•Explainable Decision Support System for industrial roundwood sorting.•System 1: SegDT - Combining SEGmentation and Decision Tree, achieving 80% accuracy.•System 2: Enhancing CNN interpretability with GradCAM.•User study comparing systems on experience and trust.•CNN recommended for automation; SegDT for human support.