Detail publikace

Performance of discrete wavelet transform lifting scheme approach

Originální název

Performance of discrete wavelet transform lifting scheme approach

Anglický název

Performance of discrete wavelet transform lifting scheme approach

Jazyk

en

Originální abstrakt

The fast lifting-based approach for the discrete wavelet transform presents a very efficient way of computing wavelet coefficients. We have successfully created a Java implementation of this coding technique for the case of CDF 9/7 filterbanks, which is a starting point of proposing lossy image compression format based on DWT, that will make use of this platform. Testing of the algorithm has confirmed it to be superior when compared to the standard convolutional approach, both in terms of computational speed and memory efficiency. The algorithm is also much simpler to implement than the standard convolutional way, because we don't need to perform manipulation on the source signal in order to archieve boundary extension and cutoff (after the transform).

Anglický abstrakt

The fast lifting-based approach for the discrete wavelet transform presents a very efficient way of computing wavelet coefficients. We have successfully created a Java implementation of this coding technique for the case of CDF 9/7 filterbanks, which is a starting point of proposing lossy image compression format based on DWT, that will make use of this platform. Testing of the algorithm has confirmed it to be superior when compared to the standard convolutional approach, both in terms of computational speed and memory efficiency. The algorithm is also much simpler to implement than the standard convolutional way, because we don't need to perform manipulation on the source signal in order to archieve boundary extension and cutoff (after the transform).

Dokumenty

BibTex


@inproceedings{BUT22897,
  author="Jan {Malý} and Pavel {Rajmic}",
  title="Performance of discrete wavelet transform lifting scheme approach",
  annote="The fast lifting-based approach for the discrete wavelet transform presents a very efficient way of computing wavelet coefficients. We have successfully created a Java implementation of this coding technique for the case of CDF 9/7 filterbanks, which is a starting point of proposing lossy image compression format based on DWT, that will make use of this platform. Testing of the algorithm has confirmed it to be superior when compared to the standard convolutional approach, both in terms of computational speed and memory efficiency. The algorithm is also much simpler to implement than the standard convolutional way, because we don't need to perform manipulation on the source signal in order to archieve boundary extension and cutoff (after the transform).",
  booktitle="Proceedings of the 12th IFIP International Conference",
  chapter="22897",
  year="2007",
  month="september",
  pages="231",
  type="conference paper"
}