ABSTRACT
Timely detection of lung cancer is important, significantly impacting patient prognosis and decreasing mortality rates. computed tomography (CT) scans have become a cornerstone in this endeavor due to their ability to provide detailed anatomical information. However, a persistent challenge in this field is striking the delicate balance between precision accuracy, and execution time during the detection process. Existing precision-focused methods often demand extensive computational resources, leading to prolonged execution times – undesirable in time-sensitive clinical scenarios. This paper introduces a groundbreaking solution by proposing a novel hybrid classification algorithm for CT image analysis. The algorithm achieves exceptional precision while substantially reducing execution times. It integrates gray-level co-occurrence matrix (GLCM) analysis into its core, efficiently identifying cancerous regions within CT scans. This approach comprises a sequential process: GLCM analysis, feature extraction, hybrid classification, algorithm training, and detection, resulting in high-precision and accurate lung cancer detection within minimal execution time. From the results, it is clear that SURF surpasses SIFT with a minimum error rate of 16.71 compared to SIFT’s 39.02. SURF also executes faster, taking 0.096 s vs. SIFT’s 3.46 s. As a result, SURF is expected to have superior recall and precision. Hence, this research addresses a critical need in the field, offering a promising pathway toward expedited, precise, and scalable lung cancer diagnosis.
