This model (Q-Eval-Score) is proposed in the paper "Q-Eval-100K: Evaluating Visual Quality and Alignment Level for Text-to-Vision Content". It is designed to comprehensively assess the quality and alignment of AI-generated visual content across both images and videos.
The model provides evaluation in four key dimensions:
Text-Image Alignment Evaluation: A model for assessing the alignment between AI-generated images and their corresponding textual descriptions.
Image Quality Evaluation: A model dedicated to evaluating the perceptual quality of AI-generated images.
Text-Video Alignment Evaluation: A model for measuring the alignment between AI-generated videos and their associated textual descriptions.
Video Quality Evaluation: A model focused on evaluating the visual quality of AI-generated videos.
