Creación de una herramienta para el diagnóstico de retinopatía diabética con técnicas de aprendizaje profundo

Deep learning techniques have had a stellar reception by the scientific community in the last decade, this, therefore, their predictive powers have been triggered by the exponential growth in computing capacity (computational power) and the massive access to large volumes of data. Thus, one of the areas where deep learning has had a greater impact is computer vision, which consists of making use of algorithms that allow recognition and classification of images and video, without the need for human intervention. Diabetic retinopathy (DR) is one of the main microvascular complications of type 1 and type 2 diabetes. In adults aged 20 to 74 years, it is the most common cause of blindness in developed countries. The prevalence of DR is strongly related to the duration of diabetes, glycemic controls, and risk factors such as chronic hyperglycemia, dyslipidemia, and hypertension. Thus, it is in our interest to make use of deep learning techniques to create a tool that allows, effectively and reliably, to make diagnoses of diabetic retinopathy autonomously (without the intervention of a specialist). In this way, it is expected to ease the burden on the Colombian health system, as well as to reduce the time for diagnosis, which is one of the crucial points to achieve effective treatment against this pathology.

• Autores: Restom Viera, José Renato
• Tipo de recurso: Trabajo de grado de pregrado
• Fecha de publicación: 2020
• Institución: Universidad de los Andes
• Repositorio: Séneca: repositorio Uniandes
• Idioma: spa
• OAI Identifier: oai:repositorio.uniandes.edu.co:1992/51492
• Acceso en línea: http://hdl.handle.net/1992/51492
• Palabra clave: Retinopatía diabética; Aprendizaje automático (Inteligencia artificial); Ingeniería
• Rights: openAccess
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/