Smoker Recognition from Lung X-ray Images using ML

Smoking is the leading cause of death in Bangladesh, accounting for one in every five fatalities, according to scientific research. A number of researchers have developed cutting-edge techniques based on Deep Learning methods to ascertain a man's smoking status through image processing. As far as we are aware, no CNN-based system can distinguish between a smoker and a non-smoker from a lung X-ray image. In this work, we offer a novel CNN-based system that can detect, in real time, with high sensitivity and specificity, whether a man smokes or not by examining lung X-ray pictures. The employed data-set is divided into two groups: smokers and non-smokers. In this study, we provide a novel ML-based system that uses images of the guy's lungs taken in real time to detect, with high sensitivity and specificity, whether or not the man smokes. Hospitals may occasionally need to utilize this method to draw blood from a nonsmoker. Additionally, men who vape are not allowed to take college admissions exams or the army selection process. The efficacy of the proposed method for Smoker and Non-Smoker prediction was evaluated and contrasted with previous CNN systems based on multiple performance metrics. The proposed technique accurately detects smokers and nonsmokers from lung X-ray pictures with 91.50% accuracy, 92% precision, and 91% recall. We are certain that real-time system performance won't be affected, even if the recommended approach was trained on a picture dataset. This device might be used in a hospital, while choosing applicants for the police or army, or for university admission.