Enhanced Lung Cancer Detection Using Advanced ROI Extraction and Deep ANN-CNN Hybrid Classifiers

Lung Cancer is a leading cause of death and can be cured if diagnosed early. For the automated detection, the system employing the image processing, artificial neural networks and convulational neural networks (ANNs & CNNs) have been developed. First the images are preprocessed using filter techniques such as noise elimination, image sharpening and segmentation which zooms into region of interests (ROI) where tumours are expected. Following preprocessing these segments of images are further analyzed in layers by CNNs to recognize significant features in layers and ANNs are used to classify the tumors as benign or malignant. This makes the approach more accurate, sensitive and specific compared to the traditional approaches of detection. These systems demonstrate higher performance using more extensive lung cancer data sets and contribute to increasing the availability of accurate early diagnosis.