We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Dimensionality Nanostructures Classification on Scanning Electron Microscopy Images Using Two-Stage Hybrid Classification Algorithm.
- Authors
Selvaraj, Rohini; Nagarajan, Sureshkumar
- Abstract
Nanoscience is increasingly focusing on artificial intelligence approaches to handle the complexities of everyday living demands. Imaging approaches for the categorization of nanomaterials with specified attributes to meet the needs of applications are of interest to researchers in both academia and industry. Machine-learning advancements have been utilized to improve computers' ability to understand scanning electron microscopy (SEM) images. The aim of the work is to classify the SEM images based on their dimensionality using a two-stage hybrid classification technique. In two-stage hybrid classification, the support vector machine (SVM) is employed in the first stage of classification. Based on the result of SVM classifier, k -nearest neighbor (k -NN) is used for final classification. The proposed classification algorithm was able to successfully classify 86.5% of test data into six categories including particles, powders, fibers, wires, thin-films and micro-electromechanical systems (MEMSs) devices. The performance of the proposed method is measured in terms of overall accuracy, class accuracy, precision, recall and f 1-score. The two-stage hybrid classification algorithm was used to classify the dimensionality of nanostructures. The SEM images are categorized into six types: 0-Dimensional (0D) items like particles and powders, 1-Dimensional (1D) objects like fibers and wires, 2-Dimensional (2D) things like a thin film and MEMS devices. In a two-stage hybrid classification, the SVM classifier is combined with the k-NN classifier. The performance of the proposed method is measured in terms of overall accuracy, class accuracy, precision, recall, and f1-score.
- Subjects
CLASSIFICATION algorithms; SCANNING electron microscopy; K-nearest neighbor classification; THIN film devices; SUPPORT vector machines; NANOSTRUCTURES; PIEZOELECTRIC thin films
- Publication
NANO, 2022, Vol 17, Issue 8, p1
- ISSN
1793-2920
- Publication type
Article
- DOI
10.1142/S1793292022500576