AI Systems for Identifying Chest and Gastrointestinal Disease

The most important by the number of fatalities and morbidity are lung cancer (153) and colon carcinoma in population.level with fourth place respectively. Simultaneous develpopment may threaten human life if they develop in the organs. If cancer is not caught early, it spreads more to both organs. Fortunately, identifying these tumors histologically is one of the linchpins in an effective therapeutic plan. Deep learning algorithms have the potential of allowing long and difficult procedures to be carried out faster, as well as more accurately due to reduced resource crystallization costs (as a result researchers can study much more patients for less cost with lesser time). Existing works have mainly focused on deep learning models, and these middleware cater to the compute intensive resources. Most of them used different DL models for diagnosis and high-dimensional feature extraction. Instead, we propose in this work a framework for early lung/ colon tumor detection using several lightweight DL models. The highest accuracy derived from this approach is 99.6%. A comparative evaluation of the proposed framework using a light weight deep learning model and other existing techniques with our results illustrate that specific types of lung cancer and colon cancer can even be detected from low feature size, lower computational cost. We also show how qualities can be normalized by transformation methods which could help in information understanding and diagnostics.