Dimensionality reduction for AI based hyperspectral image classification based on XAI

This research addresses the challenge of limited materialnrecycling in wood recycling processes by leveragingnartificial intelligence (AI)-based dimensionality reduction.nOur study explores the application of convolutional neuralnnetworks (CNNs) in multi-channel hyperspectral imagingn(HSI), extending beyond RGB channels to over 200 spectralnchannels. Dimensionality reduction within this context involvesnstreamlining the feature space for AI system trainingnand inference. Focusing on explainable AI (XAI) methods,nthis paper contributes to a broader research initiative, presentingna solution framework that enhances the sustainabilitynand efficiency of wood recycling processes.