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

Téléchargement	Télécharger l’ensemble des données partie 1 : Autoencoder-based detection of nanoplastics in biological matrices via infrared hyperspectral imaging (ML_ILIM_AE_TRAIN.IPYNB, 48 Kio) Télécharger l’ensemble des données partie 2 : Autoencoder-based detection of nanoplastics in biological matrices via infrared hyperspectral imaging (ML_ILIM_AE_EVAL.IPYNB, 57 Kio) Télécharger l’ensemble des données partie 3 : Autoencoder-based detection of nanoplastics in biological matrices via infrared hyperspectral imaging (ML_ILIM_AE_EVAL_VOILA.IPYNB, 148 Kio) Télécharger l’ensemble des données partie 4 : Autoencoder-based detection of nanoplastics in biological matrices via infrared hyperspectral imaging (ILIM_AE_ENVIRONENT-CONDA-FORGE-CUDA.YAML, 9 Kio)
DOI	Trouver le DOI : https://doi.org/10.4224/40004000
Auteur	Rechercher : Prezgot, Daniel¹Identifiant ORCID : https://orcid.org/0000-0001-6498-4422; Rechercher : Sakib, Sadman¹; Rechercher : Chen, Maohui¹; Rechercher : Pegoraro, Adrien¹Identifiant ORCID : https://orcid.org/0000-0003-3334-227X; Rechercher : Corriveau, David¹; Rechercher : Zou, Shan¹Identifiant ORCID : https://orcid.org/0000-0002-2480-6821
Affiliation	Conseil national de recherches Canada. Centre de recherche en métrologie
Format	Texte, Ensemble de données
Résumé	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.
Date de création	2026-03
Maison d’édition	National Research Council Canada
Licence	Creative Commons Attribution - Partage dans les mêmes conditions 4.0 International (CC BY-SA 4.0) https://creativecommons.org/licenses/by-sa/4.0/deed.fr
Note	Description of contents: Attached is python code in the form of Jupyter notebooks used for the training and use of models used in this work. - ML_ILIM_AE_Train.ipynb - Contains code for the training of FC AE, CNN AE, and CNN-FC AE models used for the residual anomaly detection process. - ML_ILIM_EVAL.ipynb - Contains code for applying the guided or unguided residual anomaly detection process to a target dataset using a pre-trained model. - ML_ILIM_AE_Eval_Voila.ipynb - A version of the ML_ILIM_AE_Eval which can be run as a standalone web application using Voilà. Contains UI elements for configuration of processing parameters and figure display. - ILIM_AE_environment-conda-forge-cuda.yaml - environment used for this work. Note: uses a CUDA-enabled version of PyTorch, reinstall PyTorch if not using a GPU.
Langue	anglais
Exporter la notice	Exporter en format RIS
Collection	Données de recherche du CNRC
Identificateur de l’enregistrement	224e3df6-b2e4-47f8-b498-3eee29ed9c2b
Enregistrement créé	2026-03-25
Enregistrement modifié	2026-04-09

Détails de la page

Par :

Conseil national de recherches Canada

Date de modification :: 2026-05-20