Detection of Lung Cancer in CT scans via Deep learning and Cuckoo Search Optimization and IOT

Abstract

Radiologists spend a lot of time and effort sifting through CT scans to find cancerous lung lesions. Using image processing techniques, a new diagnostic tool for medical purposes, the suggested method is applied In order to implement the proposed system, there are two steps. Optimization and deep learning are used to detect lung cancer in the early stages. In the second step, data is transferred from the MATLAB to the authorised PC using ThingSpeak. After a database search, the CT image is pre-processed using a median filter to remove background noise. Next, the Otsu-Thresholding approach segments the image together with the Cuckoo scan, and the feature extracts identify the illness size and location. Finally, the segmented image is sent to a deep learning system, which determines whether or not it contains anything normal or abnormal. ThingSpeak, MATLAB's IoT cloud, receives the final parameters. The suggested technique additionally measures and compares current system outcomes with various PSNR, correlation, precision, specificity, sensitivity, and MSE metrics.