Desarrollo de un sistema de visión artificial para la detección temprana del Alzheimer en un hospital de tercer nivel en Guayaquil

Abstract

The present project involves the development and implementation of a prototype computer vision system for the early detection of Alzheimer’s disease in a tertiary-level hospital in Guayaquil. This initiative arises from the high prevalence of the disease, delayed diagnosis, and the limited integration of emerging technologies within the Ecuadorian healthcare system. The proposal consists of designing and validating a tool that, through artificial intelligence and medical image processing (MRI), identifies early patterns of brain atrophy associated with Alzheimer’s disease. The system integrates stages such as data collection and preparation, anatomical segmentation using U-Net models, extraction of morphological features, training and validation of CNN, and multitask adaptation to identify brain regions and types of damage. Its implementation is planned in a web environment with an intuitive interface for clinical staff, pilot tests with real data from a tertiary-level institution, statistical performance analysis, and cloud deployment via Render. The system aims to optimize early diagnosis, reduce costs, and strengthen local data processing in order to drive continuous improvements and generate valuable information for future research.