A Novel Framework for Lung Nodule Classification with Low False Positive and High True Positive Detection Rate

Popuri Ramesh Babu and Dr. Inampudi Ramesh Babu

Lung cancer diagnosed as very dangerous in all disease models. One of the causes for it is that it usually is not diagnosed until it has evolved, and by that time, it becomes challenging to manage. A study on the immediate detection of lung cancer has extended over the last few years. A lot of various kinds of difficulties have addressed during this time. Multiple methods to diagnose lung cancer such as Chest Radiograph, Computed Tomography, Magnetic Resonance Imaging, etc. are provided. But, most utmost of these methods is time-consuming as well as costly. Sometimes nodules may be missed by the radiologists. It enthuses researchers to concentrate on detecting lung cancer in its beginning step to improve the possibility of endurance in a patient. An effective nodule exposure method can play a vital role in the immediate detection of lung cancer, thus increasing the Low false positive and high True positive detection rate for successful operation. In this approach, we have suggested a classification framework for Lung nodule detection. The proposed Framework includes various steps that include Image Enhancement, Image Segmentation, feature extraction(FE), followed by the employment of these features for training and testing of Enhanced Random Forest Algorithm (ERF). The accuracy achieved by ERF is the most value associated with other machine learning classifiers such as KNN, SVM, and DT. The proposed method designed to examine and select features that maximize classification results. The experimental outcomes offer that the recommended way is beneficial for reducing false-positive rates. We were able to receive an impressive sensitivity rate of 92.68%.

Volume 12 | 03-Special Issue

Pages: 1335-1344

DOI: 10.5373/JARDCS/V12SP3/20201383