FabRigo: modelo informático para el monitoreo y control de rutinas de pausas activas en trabajadores de oficina por medio del modelo posenet

Abstract

Occupational diseases are one of the biggest causes of disability in the world, these can be prevented from active pauses, however, they are not executed due to ignorance, forgetfulness or lack of motivation. This research aims to provide a mechanism that guarantees the realization of active pauses supported by a computer tool using PoseNet, a tool based on convolutional neural networks that allows estimating the key points of the human body and thus monitoring a routine of exercises related to active pause. The generated application is able to detect when the exercise is performed correctly with an error between 10^-2 and 10^-7 by using a camera like the one found on mobile devices or computers behind At Kinect, the application is capable of detecting and comparing the figure of the user to determine if he correctly performs an active pause exercise.