Analysis of Lung Cancer Prediction System using various Feature Extraction Techniques
This study investigates different computer aided diagnosis (CAD) techniques that allow detection of lung cancer through analysis of Lung computed tomography (CT) images. Various handcrafted feature schemes have been proposed for medical image classification. As far as the recent researches have been made, the primary focus has been to built an automated computer aided diagnosis (CAD) system for predicting probability of a patient, but how different features and their combination and increase the efficiency of the system. It is an attempt to understand the power of feature extraction techniques, and how various ways we can extract features from an image. We present a comparative analysis of model created for predicting the possibility of nodules being cancerous or not, using various feature techniques. The idea behind is to understand what features would be most suitable for solving medical problems. This Study will also make comparative analysis on which type of classifiers or Regressors are best suited for predicting Cancer. A comparative evaluation of both handcrafted and deep learned features for medical image classification is presented in this research. The experiments are performed on Kaggle Data Bowl Challenge 2017 Dataset.