SOA-GF: Spiral Optimized Algorithm with Gabor Filter for Automatic Lung Cancer detection and Classification

Lung cancer really has no earlier symptoms other than detecting some minute features pulmonary modules of a CT input image of a patient. The World Health Organization (WHO) report is not a good sign for doctors, radiologists, patients and the entire mankind. Computer Aided Detection (CAD) systems play a vital role in detecting these nodules and also help in minimizing the human error while processing a CT image. Many deep learning algorithms have been proposed in the recent past and produced significant results in bio-medical image processing. In this paper, we propose a Spiral Optimization Algorithm along with Gabor filter (SOA-GF) to perform feature extraction on the undertaken CT input group images. The classification was performed with Naive Bayes, Random Tree, Artificial Neural Networks (ANN) and Convolutional Neural Networks (CNN). The experimentation results have shown that the algorithm performs well will all the classification techniques of this study, yet, its combination with CNN delivered better accuracy of 87.67% compared to other classification methods. The results were presented in both tabular values and graphical representation.