Detail publikace

Concept of Variation Detector Used in Video Signal Compression Domain

Originální název

Concept of Variation Detector Used in Video Signal Compression Domain

Anglický název

Concept of Variation Detector Used in Video Signal Compression Domain

Jazyk

en

Originální abstrakt

This paper is concerned with a proportion similarity of the substance in images. In particular, we investigate the possibility of finding a cut position in a video sequence. The cut position detector is based on the data comparison when for two successive frames of the video sequence, several parameters are being evaluated. In this contribution we deal only with parameters in time domain, such as mean square and mean absolute error, image fidelity, normalized cross-correlation, Czekanowski distance and normalized distance. All functions are tested with a video sequence containing two cuts. Finally, a qualitative comparison of proposed methods is performed. For the cut detection in time domain is sufficient to use only two parameters: mean absolute error and normalized distance. Besides, for the color input data, the detector may scan only the luminance signal Y of YCbCr representation.

Anglický abstrakt

This paper is concerned with a proportion similarity of the substance in images. In particular, we investigate the possibility of finding a cut position in a video sequence. The cut position detector is based on the data comparison when for two successive frames of the video sequence, several parameters are being evaluated. In this contribution we deal only with parameters in time domain, such as mean square and mean absolute error, image fidelity, normalized cross-correlation, Czekanowski distance and normalized distance. All functions are tested with a video sequence containing two cuts. Finally, a qualitative comparison of proposed methods is performed. For the cut detection in time domain is sufficient to use only two parameters: mean absolute error and normalized distance. Besides, for the color input data, the detector may scan only the luminance signal Y of YCbCr representation.

BibTex


@article{BUT42165,
  author="Tomáš {Frýza} and Stanislav {Hanus}",
  title="Concept of Variation Detector Used in Video Signal Compression Domain",
  annote="This paper is concerned with a proportion similarity of the substance in images. In particular, we investigate the possibility of finding a cut position in a video sequence. The cut position detector is based on the data comparison when for two successive frames of the video sequence, several parameters are being evaluated. In this contribution we deal only with parameters in time domain, such as mean square and mean absolute error, image fidelity, normalized cross-correlation, Czekanowski distance and normalized distance. All functions are tested with a video sequence containing two cuts. Finally, a qualitative comparison of proposed methods is performed. For the cut detection in time domain is sufficient to use only two parameters: mean absolute error and normalized distance. Besides, for the color input data, the detector may scan only the luminance signal Y of YCbCr representation.",
  chapter="42165",
  journal="WSEAS Transactions on Circuits",
  number="9",
  volume="3",
  year="2004",
  month="october",
  pages="1776",
  type="journal article - other"
}