Description

ABSTRACT

Bacteria classification is an essential task in the medical field, for the diagnosis and treatment of various diseases. Typically, classification has been done by clinical specialists using conventional techniques, which do not rely on prediction approaches. Manual classification of bacteria is a time-consuming and challenging task that requires huge human efforts. As technology has advanced, the classification of micro-organisms has been possible with the aid of novel machine learning algorithms implemented on computers. Deep Neural Network (DNN) is one such promising technology that has been widely used for image classification. One of the variants of DNN is Convolutional Neural Network (CNN), which is an efficient technique for classification problems, has been used in this paper for bacteria classification. We have used ResNet-50 CNN model for classifying bacterial images into twenty categories, which are medically relevant. Using our approach, we could get an accuracy of 99.9% for classification. Experimental results show that our technique gives better results compared to the state-of-art approaches for bacteria classification.

INTRODUCTION

The pandemic of COVID-19 and horrific incidents like Influenza (1847-1848), Bubonic plague (1855- 1860), Cholera(1817 – 1824), and more have shown the need for better medical equipment and faster testing. More lives were lost due to a lack of proper diagnosis. Most of the pandemics were caused by microbes, especially Bacteria and Viruses. After doing research on these incidents one could realize that one of the major problems was the identification of the micro-organism family. In this paper, we have focused on the classification of one such micro-organism, i.e., bacteria, using computer-aided techniques. Typically, microbiologists use standard equipment and testing methodologies for recognizing the species of bacteria. This is a time-consuming process and requires large human effort. Also, manual classification is prone to errors. In the field of microbiology, computer-aided methods can be used in analyzing various micro-organisms and also for faster diagnosis. With the advancement of technology, image processing, and pattern recognition has gained significant improvement which resulted in high-performance classification algorithms and approaches. This requires efficient machine learning algorithms and techniques which can provide accurate results, based on the features extracted. Bacteria are micro-organisms without nucleus and their size ranges in micrometres. Classification of bacteria is not a trivial task, since its shape vary from spiral and sphere to rod. Therefore, classification is generally done based on their cell structure and components. For the past few years, machine learning algorithms are popular for image classification. Deep learning networks have become a topic of wide interest for researchers all over the world due to its high performance and accuracy. Convolutional Neural Network (CNN) is a promising deep learning approach for classification tasks and is widely used in medical imaging, cancer detection and other emerging medical applications. In this paper, we focus on classifying bacterial images using Convolutional Neural Nets. We have used two variants of ResNet CNN model ie., ResNet-34 and ResNet-50 for the classification of microscopic bacterial images

PROPOSED SYSTEM

In this paper, we focus on classifying bacterial images using Convolutional Neural Nets. We have used two variants of the ResNet CNN model ie., ResNet-34 and ResNet-50 for the classification of microscopic bacterial images

BLOCK DIAGRAM

CONCLUSION

In this work, we have presented a computer-aided approach for the classification of bacteria for the diagnosis of diseases. We have used deep learning neural network for this classification problem. For feature extraction and classification, we have used deep residual networks, i.e., different variants of ResNet CNN models like ResNet-34 and ResNet-50 models were used. The learning rate of the model has been presented and for validation, a confusion matrix also has been presented. We have classified twenty categories of bacteria with an accuracy of 99.9% using ResNet-50 CNN model. Comparisons with some of the existing works show that our approach gives better accuracy.