Autoencoder-based detection of nanoplastics in biological matrices via infrared hyperspectral imaging

This repository contains code used for autoencoder-based detection of micro- and nano-plastics (MNPs) in biological matrices. In this work, an autoencoder-based anomaly detection approach is employed to learn the biological matrix signal and then extract and highlight targeted extraneous signals from infrared spectra acquired with quantum cascade laser infrared (QCL-IR) microscopy. Residual-based anomaly mapping preserved characteristic spectral features, enabling heatmaps corresponding to known vibrational bands and visualization of localized nanoscale plastic (NP) accumulations in two- and three-dimensional cell culture models. Fully-connected (FC), convolutional neural network (CNN) and hybrid (CNN-FC) autoencoder architectures were evaluated, with FC or CNN-FC models accurately reconstructing spectra while preserving the spectral signature of the NPs.