Answers are grounded in this paper and cite their sources.
Analysis reveals critical criteria for visual quality and generation usability in AI-generated images, highlighting key user interactions.
The article explores the key criteria for evaluating visual quality and generation usability of images created using artificial intelligence technologies. The analysis includes both artistic and technical characteristics of the final output, as well as aspects of user interaction with generative models. The main groups of criteria are presented, including aesthetics, originality, composition, legibility, semantic accuracy, usability, detail, colour, artifacts, and resolution, along with generation process parameters such as prompt clarity, customisation flexibility, controllability, and response speed. The study offers a summary table illustrating the differences in perception and evaluation between users and experts. This approach provides an objective framework for analyzing the strengths and weaknesses of image generators and assessing their applicability in professional design practice.
No takes yet. Share an insight, caveat, or question.
Zausailov et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: