Inpaint Biases: A Pathway to Accurate and Unbiased Image Generation

TL;DR

This paper introduces the Inpaint Biases framework to improve the accuracy and reduce biases in text-to-image models, especially for unconventional concepts, by using user-defined masks and inpainting techniques.

Contribution

The paper presents a novel inpainting-based framework that enhances image generation accuracy and mitigates biases in text-to-image models for rare or complex concepts.

Findings

Significant improvement in image fidelity for unconventional concepts

Reduction of stereotypical biases in generated images

Enhanced creative potential of text-to-image models

Abstract

This paper examines the limitations of advanced text-to-image models in accurately rendering unconventional concepts which are scarcely represented or absent in their training datasets. We identify how these limitations not only confine the creative potential of these models but also pose risks of reinforcing stereotypes. To address these challenges, we introduce the Inpaint Biases framework, which employs user-defined masks and inpainting techniques to enhance the accuracy of image generation, particularly for novel or inaccurately rendered objects. Through experimental validation, we demonstrate how this framework significantly improves the fidelity of generated images to the user's intent, thereby expanding the models' creative capabilities and mitigating the risk of perpetuating biases. Our study contributes to the advancement of text-to-image models as unbiased, versatile tools for creative expression.