LUNG CANCER DETECTION USING RESNET50-CNN MODEL IN IMAGE PROCESSING

Abstract

Lung cancer poses a major public health concern, leading to the deaths of around one million people worldwide every year. Due to prevailing clinical cases, identifying lung cancer on chest CT scans has become crucial. Despite the necessity for accurate detection, automated testing systems fall short, and lung cancer screenings remain costly. Furthermore, utilizing numerous complex data sets in clinical settings demands substantial time and expertise. Nevertheless, dealing with the extensive and rapidly expanding cancer-related databases presents challenges in analysis, resulting in a lack of accuracy. To address this issue, we introduced a Resnet-50 Convolutional Neural Network (ResNet50-CNN) technique to enhance the accuracy of classifying lung cancer or non-cancer images. We preprocess lung cancer images using the Gaussian Mixture Model (GMM) to maximize the distance between the object and the background.We analyze the global best position and select features using an Enhanced Particle Swarm Optimization (PSO) algorithm.The proposed ResNet50-CNN model using a Deep Learning (DL) algorithm can accurately distinguish lung cancer images from non-cancer images.Furthermore, the performance evaluation of the proposed method can be analyzed in terms of accuracy, F1 score, sensitivity, and specificity. Moreover, when comparing the proposed ResNet50-CNN technique with previous methods, the accuracy is enhanced to 95.6%.