Perceptual Image Quality: Concept and Measurement
Jacques A.J. Roufs, Inst. for Perception Research, Eindhoven NL
Summary |
Relevance |
Rating |
Back to Index
Content Summary
Author distinguishes appreciation-oriented from
performance-oriented "perceptual image quality", which is "difficult
to provide a definition [of] which is precise".
The measurements that rely least on metric strength
assessments come from test subjects classifying the
degree of similarity or sensations of pairs of objects. Most examples
are concerned with "scalability": how perceived quality changes as
physical display parameters change.
Examples discussed:
- relationship between the physical parameters of luminance and
perceived brightness: little interaction between image size and
luminance
- Luminance and gamma (nonlinear luminance reproduction function)
- Sharpness (contrast) and gamma
- Perceived spatial resolution vs. image size
- Perceived quality vs. coding quantization (lossy compression)
Relevance to Multimedia
"Image quality engineering" in practice can benefit by making tradeoffs
that exploit the results of psychometric studies such as this
(e.g. choosing the appropriate transformation parameters for lossy
compression).
Rating
3 out of 5: the right idea, but nowhere near as elegant as its
digital-audio counterpart.
Armando Fox ([email protected])