Loading [a11y]/accessibility-menu.js
NRFSIM:A no-reference image blur metric based on FSIM and re-blur approach | IEEE Conference Publication | IEEE Xplore

NRFSIM:A no-reference image blur metric based on FSIM and re-blur approach


Abstract:

Image blurriness or definition is an important indicator to assess the quality of digital image. The objective no-reference image blur metric plays significant roles in t...Show More

Abstract:

Image blurriness or definition is an important indicator to assess the quality of digital image. The objective no-reference image blur metric plays significant roles in the design and optimization of visual perception-based auto-focus systems as well as image acquisition, transmission, enhancement, restoration and compression algorithms. In this paper, we proposed a novel no-reference image blur metric (NRFSIM) based on feature similarity (FSIM) index and re-blur approach. The experiments conducted on Gaussian blur images in LIVE database show that NRFSIM metric has about the same performance in single scenario but have better performance in multi-scenario compared with previous Tenengrad and Energy model. Further experiments on images obtained in industrial visual detection system indicate that NRFSIM metric can distinguish not only focal blur images but also motion blur images when focal blur and motion blur are both existing.
Date of Conference: 18-20 July 2017
Date Added to IEEE Xplore: 23 October 2017
ISBN Information:
Conference Location: Macao, China

References

References is not available for this document.