Adaptive Local Ternary Pattern on Parameter Optimized-Faster Region Convolutional Neural Network for Pulmonary Emphysema Diagnosis

Emphysema is a lung disease that occurs due to abnormal alveoli expansion. This chronic causes difficulty in breathing which can lead cancer. The progressive destruction of emphysema be assessed by Computed Tomography (CT) scans and pulmonary function tests. severity the may extend stage where one risk their life emphasizing early detection emphysema. Primary diagnosis done using spirometry CT for reducing mortality rates. Difficulties associated with different diagnostic procedures inter intra-observer variations have made blooming researches on more computer-aided techniques. paper intends develop technique improved deep learning strategy. initial process image pre-processing, performed histogram equalization median filtering. Further, Fuzzy C Means (FCM) clustering used segmentation. After segmentation, new Adaptive Local Ternary Pattern (ALTP) extracting pattern descriptor, further utilized classification. As contribution, Parameter Optimized-Faster Region Convolutional Neural Network (PO-FRCNN) developed performing diagnosis. enhancement formation classification accomplished Improved Red Deer Algorithm (IRDA), helps tune significant parameters positive influence accurateness. benchmark real-time dataset are experimentation. results show proposed method yields best result effectively diagnose when compared state-of-the-art