Covid Fungal Detection Using Deep Learning

Abstract

COVID-19, also known as 2019-nCoV, is no longer a pandemic, but a pandemic that has killed more than 651,247 people worldwide. COVID-19 currently has no specific treatment or cure, so living with this condition and its symptoms is unavoidable. This fact has made it increasingly difficult for the world\'s health care system, especially in developing lands. Despite the fact that there is no effective or proven clinical trial of an antibacterial agent, Although there is no mandatory goal of ending the COVID-19 epidemic, there are options that can help reduce the burden. Data development is not only limited health care programs but also economics. Famous learning, data mining, in-depth learning, and other practical strategies are examples. Other alternatives such as the 2019-nCoV epidemiological diagnosis and prognosis will be simplified. shrink, decision tree, vector support machine, Naive Bayes, and neutral network. Prior to the creation of the models, an analysis of the coefficient of correlation between different dependent and independent variables was performed to determine the interaction potential between each dependent element and independent data. Training data collection was used to train models 80 percent of the time, and the remaining 20% was used to test models.

Introduction

I. INTRODUCTION

Machine learning (ML) is one of the most advanced AI concepts, allowing for the development of automated, complex, and algorithmic purposes for processing biological or mathematical data of various sizes.

Machine learning algorithms can learn and refine their design based on targeted data collection, and adaptability by improving cost effectiveness or goal.

In addition, because many rural and remote areas are unable to carry out COVID-19 tests on a large, costly basis, having the same diagnostic / testing procedures that use machine-based technology to learn and use historical data will be very beneficial. It can also help to select people to be tested first.

II. METHODOLOGY

A. 6.3.Algorithm

1. CNN: Convolutional Neural Network (ConvNet / CNN) is an In-depth Learning algorithm that can capture embedded images, assign value (readable weight and bias) to the various elements / elements in an image and be able to distinguish one from the other. The initial processing required for ConvNet is very low compared to other partitioning algorithms. While in traditional methods the filters are hand-made, with adequate training, ConvNets has the ability to read these filters / symbols.
2. Layer Types

a. Input Structure: This layer contains immature image input width, height and depth

b. Convolution Layer: This layer combines the output volume by creating a computer dot product between all filters and image patch.

c. Activation Layout: This layer will use the functionality to activate object intelligence in the removal of the convolution layer. Some common unlock functions are active.

d. Combination Layer: This layer is occasionally used in covnets and its main function is to reduce the volume size which makes the calculation faster reduce memory and prevent overheating. The two most common types of binding layers are

e. Fully Combined Structure: This layer is a layer of normal neural network that takes the pre-layer input and calculates class points and produces a 1-D list of sizes equal to the number of classes.

III. RESULT

Conclusion

The discovery of COVID-19 is often associated with covid-19 symptoms, which can be detected by genetic testing and photography. Imaging tests may provide faster detection of COVID-19 and thus contribute to controlling the spread of the disease. Significant progress has been made in the in-depth classification of CNN medical images, thanks to the availability of large annotations website. CNN allows you to learn the most relevant and consistent image features in the data. However, the accuracy of the annotations data remains a major challenge in addressing the realities of COVID-19 from fungal images. Adam\'s method of editing was used to develop CNN model parameters, with cross-entropy as a function of loss. The learning rate starts at 0.001 and decreases after four times if the loss rate does not improve through the re-driving function. Models set for training 60 times. Use the stratify parameter to split our site into a training set and test set, saving part of the original site target for advanced ratings and duplicate findings.

References

[1] Stoecklin, Sibylle Bernard, Patrick Rolland, Yassoungo Silue, Alexandra Mailles, Christine Campese, Anne Simondon, Matthieu Mechain et al. \"The first cases of coronavirus 2019 (COVID-19) in France: surveillance, investigation and control measures, January 2020.\" Eurosurveillance 25, no. 6 (2020) [2] WHO WHO coronaviruses (COVID-19). Retrieved March 30, 2020 from https://www.who.int/emergencies/diseases/novel-coronavirus-2019. [3] Prasad, Sharda, Varsha Potdar, Sarah Cherian, Priya Abraham, and Atanu Basu. \"Transmission electron microscopy imaging for SARS-CoV-2.\" Indian Journal of Medical Research (2020). [4] Leppard, Gary G. In Nature Particles, pages 231-289. CRC Press, 2019.

Copyright

Copyright © 2022 Ketan Gaikwad, Prajwal Sonawane, Rutvij Choudhari. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.