GPT-4V(ision) is a Human-Aligned Evaluator for Text-to-3D Generation
This addresses the problem of scalable and adaptable evaluation for text-to-3D generation, offering a cost-effective alternative to user studies, though it is incremental as it builds on existing vision-language models.
The paper tackles the lack of reliable evaluation metrics for text-to-3D generative models by introducing an automatic, versatile, and human-aligned metric using GPT-4V, which strongly aligns with human preferences across different criteria.
Despite recent advances in text-to-3D generative methods, there is a notable absence of reliable evaluation metrics. Existing metrics usually focus on a single criterion each, such as how well the asset aligned with the input text. These metrics lack the flexibility to generalize to different evaluation criteria and might not align well with human preferences. Conducting user preference studies is an alternative that offers both adaptability and human-aligned results. User studies, however, can be very expensive to scale. This paper presents an automatic, versatile, and human-aligned evaluation metric for text-to-3D generative models. To this end, we first develop a prompt generator using GPT-4V to generate evaluating prompts, which serve as input to compare text-to-3D models. We further design a method instructing GPT-4V to compare two 3D assets according to user-defined criteria. Finally, we use these pairwise comparison results to assign these models Elo ratings. Experimental results suggest our metric strongly align with human preference across different evaluation criteria.