Lung Cancer Detection Using Regularized Extreme Learning Machine and PCA Features

Abstract

When finding abnormalities in target images is time-sensitive, as it is with many cancer tumors, image processing techniques have recently found widespread use across various medical industries to improve images for early detection and treatment stages. In the setting of cancer, where time is of the essence in detecting anomalies within medical imaging, our research takes on further urgency. In the medical field, image processing techniques have taken center stage, with the goal of improving image quality for the purpose of early identification and treatment planning. Our suggested approach incorporates multiple stages of image processing, such as feature extraction, morphological techniques, segmentation, and histogram equalization, with a focus on CT scan pictures. Finding better ways to interpret images for early detection in medical imaging is the driving force behind this research. Feature extraction, morphological algorithms, segmentation, and histogram equalization are some of the image-processing methods used in the study. In order to make the estimation process faster and more accurate, we also use Principal Component Analysis (PCA) and a Regularized Extreme Learning Machine (RELM). The suggested model performs admirably, with an accuracy of around 99.7%. When put to the test against popular models such as CNN, SVM, SVM-RBF, and RELM, the proposed method clearly comes out on top. This study's findings provide a more effective and efficient way for abnormality detection in medical imaging, which is a major advancement in the area. Early diagnosis and treatment planning in medical situations can be directly influenced by the integration of PCA and RELM, which shows promise for enhancing the speed and precision of estimation procedures