Genetic algorithms and convolutional neural networks for computer-aided diagnosis of spinal compression fractures

• Grant number: 19/01219-2
• Support type: Scholarships in Brazil - Master
• Effective date (Start): October 01, 2019
• Effective date (End): March 31, 2020
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal researcher: Renato Tinós
• Grantee: Rafael Silva Del Lama
• Home Institution: Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto (FFCLRP). Universidade de São Paulo (USP). Ribeirão Preto , SP, Brazil
• Associated research grant: 13/07375-0 - CeMEAI - Center for Mathematical Sciences Applied to Industry, AP.CEPID

Abstract

Spinal Compression Fracture (SCF) occurs when the vertebral body is collapsed. SCFs can be caused by trauma (benign SCFs) or tumors (malignant SCFs), and the investigation of the etiology of an SCF is essential for the determination of the patient's treatment. In this work, images of vertebrae obtained by Magnetic Resonance will be classified using Machine Learning. Currently, there has been a great interest in using Convolutional Neural Networks for the classification of medical images. However, such networks often require large databases that are often not available in medical applications. In addition, such networks generally do not use additional clinical information that may be important for classification. Here, different characteristics will be used for classification. The classification will be made by an Artificial Neural Network, which will have as inputs features extracted from three different sources: i) intermediate layers of Convolutional Neural Networks; ii) image preprocessing; (iii) additional clinical information. Due to the large number of available features, a Genetic Algorithm will be used for feature selection. (AU)