Non-Smokers' Lung Cancer Bio-Images Detection Using Deep Learning Approach

Lung disease can impair the respiratory system, is a foremost cause of human disease in the global space which impacts on its mortality rate. Mitigation in the detection of Histopathological bio-images (kind of a Lung cancer) is more powerful than detecting CT and X-Ray images. Screening the early Lung cancer stages in the image detection diagnosis approach would help in treatment on patients and doctor's decision strategy as well. Deep Learning (DL) techniques being emerged in the line of Medicare prediction, and its shown high level treatment efficiency and improving the care efficacy. Hence, here Convolutional Neural Networks (CNN), is an algorithm which exist sin DL domain and can be utilized in the classification of lung cancer images detection. Here, we investigated three lung cancer classes (Benign, Aden carcinomas, Squamous cell Carcinomas) - classification among Non-smokers by utilizing VGG-16-Data set. Data source is outsourced from Kaggle.com, which contains 15, 000 cases out of which 750 total cases are experimented via Jupiter Notebook-Python. Results delivered a very good achievable accuracy like 97.99% with ten iterations cycle.