Iron Oxide Nanoparticle Image Analysis Using Machine Learning Algorithms

Abstract

AbstractDigital image analysis is an effective, least time-consuming, minimal human interventional, optimal resource-oriented technique used in the last few decades. Computerized image analysis has been of great aid to young scientists and researchers in setting a benchmark in various areas. The major applications of Iron Oxide Nanoparticles are such as; curing cancer, drug delivery systems, antifungal activities, imaging, and cellular labeling. In this study, the Transmission Electron Microscopy(TEM) images (a, b, c, d, and e) of Iron Oxide Nanoparticles are synthesized using chemical and biological methods at 300 ℃ and 500 ℃ temperature. Images a, b, c, and d are prepared using chemical methods and image e is prepared by biological method with Iron Nitrate and Rudanti fruit plant extract. The motive behind the present work is that the traditional characterization techniques are time-consuming and are not economically cost-effective. Hence, an effort is made to automate a tool that can determine the size of TEM images of Iron Oxide Nanoparticles. The Gaussian Mixture Model - Expectancy Maximization (GMM-EM) segmentation technique is applied and diverse features, namely; size (area), perimeter, major axis, minor axis, porosity, circularity, interparticle distance, and average area are computed from TEM images of Iron Oxide Nanoparticles. It is observed that the area-wise percentage of images a, b, c, and d is 43.39%, 62.06%, 57.14%, and 70.00%, respectively, which are synthesized by using the chemical method and the area-wise percentage of image e is 64.28%, by using the biological method. The biological synthesization method is universally accepted as it uses no toxic chemicals, is cost-effective, and is environmentally friendly. The proposed results are analyzed and compared with manual results obtained from the chemical experts and are found to be a good performance.KeywordsTEM. iron oxideNanoparticlesChemicalBiologicalRudantiSegmentationImage analysisGMM-EM