Vector image recognition has dispersed into the practical systems of the Internet of Things (IOT) and research settings to further elucidate segmented properties of an image. The mechanisms of Artificial Intelligence (AI) protocols in image recognition have advanced in further monitoring irregular geometric anomalies of a picture. This universal process of image analysis has become increasingly relevant in complex medical diagnostics. Machine learning AI diagnostic applications have advanced in limited single output recognition of internal and superficial disease types. However, there has been negligible development in real time image analysis towards revealing and diagnosing critical areas of the image. This project uses a novel presymptomatic heat mapping-image recognition technique to detect and diagnose medically concerning areas of brain PET scan Amyloid-beta/Tau protein deposits. Moreover, the heat map assessment indicates loci of the brain prone to Alzheimer’s disease based on low threshold application programming interface (API) confidence scores. Full image analysis of open-source high-frequency amyloid/tau PET scans were collected into positive, moderate, and negative classifiers containing 60 absolute JPG files. The trial resources IBM’S Bluemix API to process images into a trained neural network module. Five testing phases in each pre-processing stage were assessed by average threshold confidence scores (~0.15). Highest output accuracy through heat mapping was revealed through edge-based detection and image tile separation with nearly 95% accuracy-confidence.