We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
GoogLeNet/DenseNet-201 to classify near-infrared (NIR) spectrum graphs for cancer diagnosis – using pretrained image networks for medical spectroscopy
Summary
This study compared pretrained deep learning image classification networks—GoogLeNet and DenseNet-201—with traditional machine learning methods for classifying near-infrared spectra of cancerous and non-cancerous tissue, a methodology relevant to spectroscopic identification of microplastics in biological samples.
Abstract The study compares sensitivity/specificity of classification by pretrained image networks and traditional Machine Learning (ML) methods. One hundred seven spectra each of benign skin conditions actinic keratosis (ACK) and seborrheic keratosis (SEK), and skin cancer basal cell carcinoma (BCC) were downloaded from a public database. Eighty spectra per group were used for training and twenty-seven spectra per group for testing. In the first classification strategy, spectrum intensity values were used as input for Linear Discriminant Analysis (LDA), K-Nearest Neighbors (KNN), Decision Tree (DT), TreeBagger, Ensemble method, Naïve Bayes, Support Vector Machine (SVM), and Artificial Neural Network (ANN). The second strategy involved using spectrum graphs saved as images to train GoogLeNet, Places-365 GoogLeNet, ResNet-50, Inception-V3, DenseNet-201, and NasNetMobile. Strategy 2 yielded better sensitivity/specificity – 0.7/ 0.91 (ACK), 0.7/0.83 (BCC), and 0.63/0.85 (SEK) compared to strategy 1–0.52/0.94 (ACK), 0.7/0.8 (BCC), and 0.5/0.8 (SEK). Grad-CAM mapping suggested that 1100–1200,1350–1450, and 1600–1700 1/cm to be responsible for classification by strategy 2. When these regions plotted as subplots and saved as images were used for training using strategy 2, sensitivity for BCC increases to 0.78. Results suggest using pretrained image networks to classify spectra may yield better results, give a visual understanding of the basis of classification, and provide means to improve classification further.
Sign in to start a discussion.
More Papers Like This
Deep convolutional neural networks for aged microplastics identification by Fourier transform infrared spectra classification
This study developed a deep learning model using convolutional neural networks to automatically identify aged microplastics from their infrared spectra. Aging changes the chemical signature of plastics, making them harder to identify with conventional spectral databases. The AI approach achieved high accuracy and could significantly speed up the analysis of environmental samples where weathered microplastics are the norm.
Deep learning for chemometric analysis of plastic spectral data from infrared and Raman databases
A novel deep learning architecture called PolymerSpectraDecisionNet was trained to identify common recyclable plastics from infrared and Raman spectral databases. The model outperformed conventional chemometric methods for polymer classification and was designed to handle real-world spectral variability relevant to the plastics recycling industry.
Identification of Polymers with a Small Data Set of Mid-infrared Spectra: A Comparison between Machine Learning and Deep Learning Models
Researchers compared multiple machine learning and deep learning models for identifying polymer types from mid-infrared spectral data using a small reference dataset, finding that certain deep learning architectures outperformed traditional methods even with limited training examples, supporting automated microplastic identification.
Spectral Classification of Large-Scale Blended (Micro)Plastics Using FT-IR Raw Spectra and Image-Based Machine Learning
Researchers developed and compared four machine learning classifiers for identifying microplastic types from Fourier transform infrared spectroscopy data using large-scale blended plastic datasets. The study found that a 1D convolutional neural network achieved the best overall accuracy at over 97%, outperforming decision tree and random forest models, offering a scalable alternative to traditional library-search methods for microplastic identification.
A Comparative Study of Machine Learning and Deep Learning Models for Microplastic Classification using FTIR Spectra
Researchers compared machine learning and deep learning models for classifying microplastics using FTIR spectra, evaluating multiple algorithmic approaches against standardised spectral datasets. The study assessed classification accuracy and computational efficiency, identifying which model architectures best discriminate between polymer types in environmental microplastic samples.