Assistive Technology through Internet of Things and Edge Computing
M Villadangos Carrizo Jose, Orlewilson Bentes Maia, Horácio A.B.F. Oliveira, Eduardo Souto, Raimundo Barreto · 2019
This paper proposes a framework that applies computer vision and machine learning techniques, through an Internet of Things (IoT) network, with the use of cloud computing for capacity increase. Images are captured by an IoT device and sent for an edge element (IoT node), which processes them, identifies objects, computes distances, and, ultimately, converts that information into audible commands, in order to provide guidance for visually-impaired people. Unrecognized objects are sent to a cloud service, in order to provide network retraining and, consequently, system evolution. With the goal of validation, the proposed framework was implemented, using known and proven technologies, such as Raspberry Pi, Esp32-CAM, OpenCv, Tensorflow, Google-TTS and Python. Finally, experimental results demonstrated its feasibility and effectiveness.