Classificação de imagens de espiral de indivíduos saudáveis e com a doença de Parkinson por meio de redes neurais convolucionais

One of the diseases that affect the central nervous system is Parkinson's disease (PD), and it is present in about 1% of the world population over 65 years old, and still remains incurable. One of the cardinal signs of PD is tremor. A possible way to evaluate tremor is through the severity scales, which use drawings, such as waves, circles and spirals. Subsequently, each drawing can be assessed in order to verify the severity of the motor sign. This work aims to contribute to the assessment of the evaluation of tremor by using images digitized from hand drawings. A set of images obtained from healthy and PD participants was classified by a convolutional neural network. The Keras API-encoded network achieved an average accuracy of 76.3% in the discrimination of healthy individuals from those suffering from P