Subjective Quality Assessment of Underwater Images Enhanced Using Deep Learning Models

This study is part of a master’s thesis entitled “Subjective and objective quality assessment of underwater images enhanced using deep learning models” by student Stella Palazari at the University North in Varaždin.

All collected data will be anonymized.

Steps

You will first see 5 training images.
These are only for practice. Your ratings will not be recorded.
You will then evaluate 132 test images.
The images come from four underwater image enhancement algorithms. Some test items also include the original (pre-enhancement) images and, in some cases, quasi-reference images.
For each image, please provide your quality rating according to the on-screen instructions.

What does “quasi-reference” mean?

In this context, a quasi-reference is an assumed high-quality version of an image used for comparison when a true ground-truth reference is not available.

Subjective Evaluation of the Quality of Underwater Images Enhanced Using Deep Learning Models

Welcome!

Please fill in the basic information before starting the study.

Name * Surname * Age (years) * E-mail * Gender * Normal or corrected-to-normal vision acuity and normal color vision? * Consent to use collected data anonymized for future research? * Illumination type * Monitor resolution? * Monitor or laptop name *