Machine Learning Approach for Nano-Magnetic $\text{Fe}_{2}\mathrm{O}_{3}$ Particle Classification

In this work is presented a machine learning approach for nano- magnetic particle classification. Method is applied by convolutional neuron network (CNN), performed over a large set of TEM image data. Classification process include $\text{Fe}2\mathrm{O}_{3}$ particle size, shape, volume and number recognition. CNN training and classification accuracy is verified by Sequential Pattern Search method and Contour tracing method.