Prediction of Future Immediate Actions

• Mehmet Ozturk ^[1]
  1. [1] Departamento de Informática y Automática, Facultad de Ciencias, Salamanca
• Localización: Avances en Informática y Automática. Duodécimo Workshop / coord. por André Filipe Sales Mendes, 2020, ISBN 978-84-09-10312-6, págs. 125-137
• Idioma: inglés
• Enlaces
  • Texto Completo Libro
• Resumen

  • Human action recognition and prediction are two of the important areas in computer vision. The researches and outputs are being used in various contexts like security surveillance systems, robotics and human-computer interaction. While human action recognition is a common research area, the prediction of human actions is still an area that needs more attention and research. The work in this paper focuses to design an easy-to-implement intelligent system to recognize humans/objects by using various basic object detection algorithms and predict their immediate future actions by training a model in neural networks as one of the main machine learning approaches. The system will predict the immediate future location of a moving object/human. The experiments that have been done are also provided step by step in this paper. The pieces of codes that have been used and the results of the experiments are provided as well