April 2011
Volume 52, Issue 14
Free
ARVO Annual Meeting Abstract  |   April 2011
Qualitative Evaluation of JPEG Artifacts Using Event-Related Potentials in EEG Experiments
Author Affiliations & Notes
  • Lea Lindemann
    Computer Graphics Lab, TU Braunschweig, Braunschweig, Germany
  • Marcus Magnor
    Computer Graphics Lab, TU Braunschweig, Braunschweig, Germany
  • Footnotes
    Commercial Relationships  Lea Lindemann, None; Marcus Magnor, None
  • Footnotes
    Support  ERC No. 256941 "Reality CG"
Investigative Ophthalmology & Visual Science April 2011, Vol.52, 6099. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Lea Lindemann, Marcus Magnor; Qualitative Evaluation of JPEG Artifacts Using Event-Related Potentials in EEG Experiments. Invest. Ophthalmol. Vis. Sci. 2011;52(14):6099.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract
 
Purpose:
 

The goal of our work is to find out if electroencephalography can be used to assess perceived image quality, as compared to user studies. As a first step, we here investigate if the reaction to compression artifacts introduced by JPEG can be detected, or even quantified, by comparing event-related potentials (ERP) for varying compression.

 
Methods:
 

EEGs were recorded from 10 healthy participants. A stimulus consisted of the uncompressed image (S1) that was displayed for 500ms, directly followed by a compressed version of the image (S2) that was also shown for 500ms. The test was performed for 3 different images, each in 7 different compression levels (from no to high compression ratios). The participants responded with a button press after the image disappeared, indicating whether they had detected a difference between the images or not. The recorded trials were sorted by stimulus and response.

 
Results:
 

There is a significant difference between the ERPs after S2 appeared depending on whether the change was detected or not, in addition, the ERPs differ depending onvary considerably with JPEG compression ratio. This effect was best observed as an oscillation at occipital sites for which the ERPs are shown, the reaction to the highest compression is very similar for all 3 test images. The change of the ERP due to decreasing compression depends on how easy the artifacts were to detect, i.e., the ERPs change faster for images in which artifacts are harder to detect.

 
Conclusions:
 

We show that the presence of JPEG artifacts elicits an ERP when the uncompressed image is followed by a JPEG-compressed version. Form and amplitude of the ERP vary with compression ratio. These first results show promise that image quality may be objectively quantified using EEG measurements.  

 
Keywords: image processing • perception • electrophysiology: non-clinical 
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×