Tese

Modelagem da perda de qualidade de videos H.264 em redes sem fio considerando perdas de PSNR e de frames

Multimedia applications have been growing in recent years; new consumptions like online games, video conference, video on demand and IP telephony are some of these. However, there is a greater prominence in the search related to videos and streaming, currently in high resolutions and mostly traff...

ver descrição completa

Autor principal: CARMONA, João Victor Costa
Grau: Tese
Idioma: por
Publicado em: Universidade Federal do Pará 2019
Assuntos:
Acesso em linha: http://repositorio.ufpa.br/jspui/handle/2011/11929
Resumo:
Multimedia applications have been growing in recent years; new consumptions like online games, video conference, video on demand and IP telephony are some of these. However, there is a greater prominence in the search related to videos and streaming, currently in high resolutions and mostly traffic over wireless networks, mainly due to the proliferation of mobile devices and significant increase of access networks, which make it more comfortable. Providing this information is easy. Thus, as an immediate consequence of this type of flow, there is a need for investments in techniques and mechanisms that provide the end user with the desired quality and satisfaction in the face of high definition content. This work aims to perform the modeling of video quality loss by analyzing their performance in various resolutions, specifically standards in HD and UHD, at 720p, 1080p and 2160p. In this sense, applying a correlation investigation between the metrics extracted from the videos, using Pearson’s correlation coefficient, and fundamentals of the area in question. Also proposing equations for quality loss modeling, based on analysis of metrics associated with packet loss, in which at the end of the study and according to notes made throughout the text, we used for the general modeling equation, the parameters of Loss of PSNR and Loss of Total Frames. The result obtained shows maximum values of RMSE and Standard Deviation of 0.793 dB and 0.810 dB, respectively, making the developed model very good for the tested video set and its resolutions.