If at First You Don't Succeed, Try, Try Again: Faithful Diffusion-based Text-to-Image Generation by Selection

TL;DR

This paper demonstrates that large diffusion-based text-to-image models are inherently more faithful to prompts than previously thought and proposes a candidate selection approach to improve faithfulness efficiently.

Contribution

It introduces a simple candidate selection pipeline that enhances faithfulness in T2I models without retraining or complex modifications.

Findings

Outperforms post-hoc methods in faithfulness metrics

Achieves comparable or lower computational cost

Validated through quantitative and user studies

Abstract

Despite their impressive capabilities, diffusion-based text-to-image (T2I) models can lack faithfulness to the text prompt, where generated images may not contain all the mentioned objects, attributes or relations. To alleviate these issues, recent works proposed post-hoc methods to improve model faithfulness without costly retraining, by modifying how the model utilizes the input prompt. In this work, we take a step back and show that large T2I diffusion models are more faithful than usually assumed, and can generate images faithful to even complex prompts without the need to manipulate the generative process. Based on that, we show how faithfulness can be simply treated as a candidate selection problem instead, and introduce a straightforward pipeline that generates candidate images for a text prompt and picks the best one according to an automatic scoring system that can leverage already existing T2I evaluation metrics. Quantitative comparisons alongside user studies on diverse benchmarks show consistently improved faithfulness over post-hoc enhancement methods, with comparable or lower computational cost. Code is available at \url{https://github.com/ExplainableML/ImageSelect}.