Desarrollo de un sistema de detección de tumores cerebrales con imágenes de resonancia magnética mediante técnicas de inteligencia artificial.

This research addresses the development of a brain tumor detection system using AI techniques applied to magnetic resonance imaging (MRI). Brain tumors represent a medical challenge due to their complexity and location within the brain tissue, in addition to the fact that manual detection of tumors requires specialized professionals and is susceptible to subjectivities of the specialist. Therefore, the proposed system is based on the use of two Deep Learning algorithms for the detection and classification of brain tumors that takes advantage of AI advances to create a tool to support healthcare professionalsIn the first stage, a neural network based on the pre-trained NasNetLarge model is employed to identify the presence of a brain tumor in the MRI image with 99% accuracy. If a tumor is detected, the second stage focuses on classifying the tumor type using a neural network based on the MobileNetV2 architecture, adapted for the classification of three tumor types: glioma, meningioma, and pituitary. This stage has an accuracy of 95%. In addition, a segmentation stage is incorporated to determine the location of the tumor within the brain tissue, which is essential for personalized treatment planning. Keywords: neural networks, image processing, transfer learning, brain tumors, magnetic resonance imaging, artificial intelligence

• 第一著者: Ramirez Ochoa, Dennis Alejandro (author)
• フォーマット: bachelorThesis
• 言語: spa
• 出版事項: 2023
• 主題: REDES NEURONALES; PROCESAMIENTO DE IMÁGENES; TRANSFERENCIA DE APRENDIZAJE; TUMORES CEREBRALES; RESONANCIA MAGNÉTICA; INTELIGENCIA ARTIFICIAL
• オンライン･アクセス: https://dspace.unl.edu.ec/jspui/handle/123456789/28654