An Efficient Nodule Characterization Framework using Gabor Filter and Shape Features

Lung cancer accounts for a large percentage of cancer-related fatalities each year, making it one of the most common and deadliest cancer types in the world. Lung cancer early detection and characterization entails detecting the disease when it is more curable and has a better probability of being successfully treated. In order to characterize lung nodules, a new completely automated machine learning (ML) framework has been created in this work by introducing Gabor filter-based textural features combined with shape features. Using the publicly available LIDC-IDRI dataset for lung nodule characterization, the suggested framework achieved the highest performance indices of $\mathbf{9 0. 4 8 \%}$ sensitivity, 100% specificity, and 95% accuracy.