RIT Color Science Seminar
In mid-July, Samuel Morillas Gómez gave a second invited talk at the Rochester Institute of Technology (RIT) — a place of special significance in his career, as he spent the academic year 2016–2017 there refining his expertise in visual image processing and fuzzy systems.
Returning to RIT seven years later to present ongoing research was both a scientific and personal milestone.
USTRESS: Measuring Fidelity with Observer Variability
The talk, “USTRESS: Measuring fidelity between psychophysical and computed data taking into account observers’ variability”, introduced a novel metric for image quality assessment that goes beyond traditional correlation coefficients like Pearson or Spearman.
By incorporating variability among human observers, USTRESS improves the agreement evaluation between computed image quality values and psychophysical scores. The method builds on the well-known STRESS metric but adds a layer of statistical depth using the covariance structure of the psychophysical data.
🧪 The presentation included both synthetic experiments and real-world tests using the CID:IQ dataset, showing how incorporating perceptual uncertainty can yield more robust and statistically significant results.
If you want to know more about it, contact us.
