Video on demand service based on inference emotions user

Abstract

Video traffic on networks increases exponentially, and thus the amount of time that should be used browsing content catalogs. Therefore, systems are needed video on demand [VoD] taking into account the emotions as a parameter for fast access to content. This paper presents the design and implementation of a VoD service based on emotions, whose main components are: the musical content catalog forming and hardware-software system that allows you to set the level of mental stress and inference of emotions of the consumer, while it interacts with the system. The final product was tested for efficiency and stress, with satisfactory results: the time spent by the web server with 200 sequential connections, ranged from 0.050 to 0.675 seconds and between 0.030 and 0.675 seconds when they are simultaneous. It also managed to respond adequately to 20,000 sequential connections, with response times of less than 1 to 36 seconds, and withstand, without collapsing, 18,000 concurrent connections, with response times between 7 and 62 seconds. The project provides an open source service that raises the groundwork for future projects.