skip to main content
Idioma:

Video Synchronization With Bit-Rate Signals and Correntropy Function

Pereira, Igor ; Silveira, Luiz F ; Gonçalves, Luiz

Sensors (Basel, Switzerland), 2017-09, Vol.17 (9), p.2021 [Periódico revisado por pares]

Switzerland: MDPI AG

Texto completo disponível

Citações Citado por
  • Título:
    Video Synchronization With Bit-Rate Signals and Correntropy Function
  • Autor: Pereira, Igor ; Silveira, Luiz F ; Gonçalves, Luiz
  • Assuntos: Access control ; Accuracy ; Algorithms ; Cables ; Cameras ; Computation ; correntropy ; Crystal oscillators ; Data transmission ; Methods ; Multimedia ; Sensors ; Spacetime ; State of the art ; Strategy ; Streams ; Synchronism ; variable bit-rate ; Viability ; Video data ; video synchronization ; Wireless networks
  • É parte de: Sensors (Basel, Switzerland), 2017-09, Vol.17 (9), p.2021
  • Notas: ObjectType-Article-1
    SourceType-Scholarly Journals-1
    ObjectType-Feature-2
    content type line 23
  • Descrição: We propose an approach for the synchronization of video streams using correntropy. Essentially, the time offset is calculated on the basis of the instantaneous transfer rates of the video streams that are extracted in the form of a univariate signal known as variable bit-rate (VBR). The state-of-the-art approach uses a window segmentation strategy that is based on consensual zero-mean normalized cross-correlation (ZNCC). This strategy has an elevated computational complexity, making its application to synchronizing online data streaming difficult. Hence, our proposal uses a different window strategy that, together with the correntropy function, allows the synchronization to be performed for online applications. This provides equivalent synchronization scores with a rapid offset determination as the streams come into the system. The efficiency of our approach has been verified through experiments that demonstrate its viability with values that are as precise as those obtained by ZNCC. The proposed approach scored 81 % in time reference classification against the equivalent 81 % of the state-of-the-art approach, requiring much less computational power.
  • Editor: Switzerland: MDPI AG
  • Idioma: Inglês
Links
Voltar para lista de resultados

  • Realização: ABCD-USP USP   Logos de Redes Sociais: Freepik

Buscando em bases de dados remotas. Favor aguardar.