International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 12 Issue: 05 | May 2025
p-ISSN: 2395-0072
www.irjet.net
DETECTION AND ANALYSIS OF LUNG CANCER TUMORS USING ML TECHNIQUES P.S.Mayura Veena1, M.Chandrika2, Akshaya.T3, S.Karthik4, Ch.Harshavardhan5 1Assistant Professor, Department of ECE, Anil Neerukonda Institute of Technology and Sciences, Andhra Pradesh,
India
2UG student, Department of ECE, Anil Neerukonda Institute of Technology and Sciences, Andhra Pradesh, India 3UG student, Department of ECE, Anil Neerukonda Institute of Technology and Sciences, Andhra Pradesh, India 4UG student, Department of ECE, Anil Neerukonda Institute of Technology and Sciences, Andhra Pradesh, India 5UG student, Department of ECE, Anil Neerukonda Institute of Technology and Sciences, Andhra Pradesh, India
---------------------------------------------------------------------***--------------------------------------------------------------------arises from epithelial cells; squamous cell carcinoma, commonly located near the central bronchi; and large cell carcinoma, which typically occurs in any lung region and is characterized by rapid growth and metastasis.
Abstract- Lung cancer is currently the most frequently
diagnosed major cancer and the most common cause of cancer mortality worldwide. Early and accurate detection results in more effective treatment options, thereby improving the patient outcomes. This study presents an automated classification model for lung cancer using machine learning techniques applied to CT scan images. At first, the dataset is divided into training , validation, and testing sets. The process begins with image pre-processing, followed by feature extraction using Histogram of Oriented Gradients (HOG) and classification using machine learning classifiers- Support Vector Machine (SVM), logistic regression, and Naive Bayes. Finally, the models are evaluated using various metrics like accuracy, precision, recall, and F1-score. Among the models, SVM achieved the highest training accuracy of 96.41%. Based on the classification results, only the images that are predicted as cancerous are passed on for segmentation, thereby reducing the computational load. The proposed model demonstrates its strong ability to assist in early detection and analysis of lung cancer tumors.
Traditional diagnostic methods like analysis of lung CT scans by radiologists require expertise and may be limited by variability and subjectivity in interpretation. Earlystage nodules, i.e., subtle nodules, may be misclassified or overlooked, thereby leading to suboptimal treatment. Therefore, to avoid these limitations, an automated diagnostic tool that supports early and accurate detection of lung cancer is developed. Machine learning provides a powerful alternative tool enabling automated, objective analysis of complex imaging data. In this study, three ML algorithms, like support vector machine (SVM), Logistic Regression, and Naive Bayes, are used for the multiclass classification of NSCLC subtypes using CT scan images.In examining the structural and textual features of lung tissues, the HOG method is employed.
Key words- Machine learning (ML), Histogram of Oriented Gradients (HOG), Naive Bayes, Support Vector Machine (SVM), Logistic Regression.
Every algorithm offers specific benefits: SVM is suited for both linearly separable data and high dimensional spaces; logistic regression is easy to understand and offers a clear rationale; and Naive Bayes is commonly used because it is simple and computationally efficient. Using three ML models, their accuracy, precision, recall, and F1-score are used to evaluate and compare the algorithms. In the proposed diagnostic model, classification is used for initial lung cancer screening. If a CT image is classified as one of the subtypes of lung cancer by the model, it goes next to the segmentation of cancerous tumor nodule nodules. This stepwise strategy lessens computational demand, avoids superfluous segmentation, and maximizes resources, improving overall effectiveness and efficiency.
1.INTRODUCTION Cancer is one of the main causes of non-accidental deaths, with lung cancer being its main contributor. Early detection enables timely and effective treatment, significantly improving the patient outcomes. The uncontrollable growth of abnormal cells forming malignant tumors leads to lung cancer. Lung cancer also has the ability to spread to the adjacent tissues and organs. Lung cancer is broadly classified into small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC). This study focuses exclusively on NSCLC, which accounts for nearly 85% of all lung cancer cases. NSCLC is categorized into three major subtypes: adenocarcinoma, majorly found in the outer regions of the lungs, which
© 2025, IRJET
|
Impact Factor value: 8.315
2.LITERATURE SURVEY As noted earlier, lung cancer is one of the most common causes of cancer death all over the world which requires timely detection and classification. There is often a
|
ISO 9001:2008 Certified Journal
|
Page 38