Sparsity-driven reconstruction of ℓ∞-decoded images

Yuanman Li, Jiantao Zhou

Abstract

In this paper, we propose a sparsity-driven restoration technique to improve the coding performance of the ℓ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">∞</sub> -decoded images. This is achieved by incorporating a ℓ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub> minimization term into a ℓ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> optimization framework, where the weighting vectors balancing the relative contribution of each term are appropriately determined. The ℓ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">∞</sub> constraints inherent to ℓ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">∞</sub> -constrained predictive coding are also included to narrow the solution space, leading to more accurate estimation. Experimental results show that our proposed scheme significantly improves the ℓ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> performance of the ℓ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">∞</sub> -decoded images, while still preserving a tight error bound on every single pixel. In addition, when comparing with the existing scheme of restoring the ℓ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">∞</sub> -decoded images, the PSNR gain can be up to 1 dB.
BibTeX
@inproceedings{li2014sparsity,
  author = {Yuanman Li and Jiantao Zhou},
  title = {{Sparsity-driven reconstruction of ℓ∞-decoded images}},
  booktitle = {ICIP},
  year = {2014},
  url = {https://openreview.net/forum?id=bgxbEpA2iXe},
}

Loading related content...

Figure