resnet101  (MathWorks Inc)

Journal: Scientific Reports

Article Title: Effective data selection via deep learning processes and corresponding learning strategies in ultrasound image classification

doi: 10.1038/s41598-025-00416-5

Figure Lengend Snippet: Grad-CAM visualization for ResNet101 applied to six ultrasound images (three benign and three malignant). The figure is structured in a 3 × 6 format, where the first column contains benign images, the second column shows the corresponding Grad-CAM results from the original network, and the third column presents the Grad-CAM results from the True network. Similarly, the fourth column contains malignant images, followed by Grad-CAM results from the original network in the fifth column and those from the True network in the sixth column. The color scale from red to blue represents the importance of different image regions, where red indicates the most critical areas used for classification. The selected images include cases where the original network misclassified or had a small probability difference, whereas the True network provided correct classification with a higher probability difference.

Article Snippet: When training ResNet101 in MATLAB, the input image size is \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:224\times\:224$$\end{document} pixels, with the minibatch size is 100, and the detailed conditions of the Adam optimizer are the gradient decay factor of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:0.9$$\end{document} , the squared gradient decay factor of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:0.999$$\end{document} , and the epsilon of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{10}^{-8}$$\end{document} .

Techniques:

Journal: BMC Medical Informatics and Decision Making

Article Title: Deep learning for the classification of atrial fibrillation using wavelet transform-based visual images

doi: 10.1186/s12911-025-02872-5

Figure Lengend Snippet: AF classification performance on train, validation, and test datasets using ResNet101 models

Article Snippet: The pretrained CNN: ResNet101, selected from Matlab's 2022a Pretrained CNN document [ ], was employed for our research.

Techniques: Biomarker Discovery

Journal: PLOS ONE

Article Title: Accelerating antimicrobial peptide design: Leveraging deep learning for rapid discovery

doi: 10.1371/journal.pone.0315477

Figure Lengend Snippet: The architecture of the ResNet101.

Article Snippet: This paper utilized the ResNet101 model already incorporated in MATLAB® version 2022.

Techniques: