Deep Learning-Based Microorganisms Classification Systems Using Microscopic Images

Abstract

With enormous amounts of microscopic images created in the research of microbiology, conventional methods of computation have turned more and more difficult to handle. On the contrary, deep learning models are inclined to hold outstanding performance in speed and accuracy. Recently, the microbiologist community embraced deep learning models, thus leading to the appearance of new applications with unprecedented discoveries and perspectives in the research of microbiology. In this paper, various microorganisms classification systems are implemented using various types of microscopic image datasets, including parasites, fungi, bacteria, and viruses. The backbone of these systems is pre-trained convolutional neural networks (CNNs) for detecting microorganisms. This study conducted an exhaustive analysis of various pre-trained CNNs models in the field of microorganism classification,as well as their experimental design and validation and the future extent to which they will present profound perception for the researcher’s active in this field.