Predicting the Patient’s Severity Using Machine Learning Applied to Lungs MRI Images

Abstract

Lung cancer has repeatedly shown itself to be one of the most fatal illnesses in the history of mankind. Additionally, it is among the most prevalent and deadly between all cancers. Lung cancer cases are rising quickly. In India, there are roughly 70,000 instances per year. Since the condition tends to be asymptomatic in its early stages, it is almost impossible to identify. Because of this, early cancer identification is crucial to preserving lives. A patient may have a better chance of recovery and cure with an early diagnosis. Effective cancer detection is greatly aided by technology. On the basis of their findings, numerous researchers have suggested various methodologies. Recently, various computer-aided diagnostic (CAD) methodologies and systems have been proposed, developed, and launched in an effort to employ computer technology to address this challenge. There are multiple ways using image processing-based methods to forecast the malignancy level for cancer, and those systems use a variety of machine learning techniques in addition to deep learning techniques, like ResNet 50 and DenseNet169 got accuracy 98.69 of ResNet 50 and 99.67 of DenseNet169. The purpose of this study is to list, debate, contrast, and analyse several methods of image segmentation, feature extraction, and other methodologies to classify and identify lung cancer in its early stages.