High Resolution Image Upsampler for Lung Adenocarcinoma diagnosis

Lung Adenocarcinoma is a kind of NSCLC (non-small cell lung cancer) which is the most occurring form of lung cancer. It progresses in the epithelial cells that coat the Alveoli sacs of the lungs. Adenocarcinoma refers to cancer formed in mucus secreting cells, thus giving it its name. Patients suffering from this disease shows a combination of symptoms such as pneumonia, coughing up blood, fatigue and weight loss which covers a wide range of unrelated manifestation leading the disease to go unnoticed and often causing fatality. Just like any other cancer, Lung Adenocarcinoma desires early and accurate diagnosis. Current methods include various forms of imaging techniques such as PET scan, biopsy procedures and genetic testing. Diagnosis is followed by a plethora of treatment options such as Chemotherapy, Targeted therapy and Immunotherapy. Despite such efficacious treatment techniques, their effect entirely relies on the stage of diagnosis and its accuracy. Keeping such dearth in mind, we put forward a high resolution upsampled image generator which is packed with combinations of advanced neural network incorporating Machine learning model and techniques such as GANs (Generative Adversarial Network). This platform utilizes pathological data from several trusted sources to develop the AI model which as a result will increase the accuracy of biopsy diagnosis. Trained on the dataset that was carefully acquired and the initial 750 samples were augmented generating the final dataset of 4500+ image samples using the ‘Augmentor’ package in python, the model delivers superior results than current diagnosis methods while simultaneously improving biopsy performance. It guarantees model stability to variations which is achieved through minimizing loss function and parameter tuning while enhancing computational efficiency. This development may eventually result in filling the gap in current diagnosis methods by improving its results manifold.