ViPer: Visual Personalization of Generative Models via Individual Preference Learning

TL;DR

ViPer introduces a method to personalize generative image models by learning individual preferences through user comments and guiding image generation accordingly, improving alignment with personal tastes.

Contribution

The paper presents a novel approach that captures user preferences via comments and uses language models to guide personalized image generation, reducing manual prompt engineering.

Findings

Generated images align better with individual preferences.

User satisfaction with personalized images increases.

Method outperforms baseline in preference alignment.

Abstract

Different users find different images generated for the same prompt desirable. This gives rise to personalized image generation which involves creating images aligned with an individual's visual preference. Current generative models are, however, unpersonalized, as they are tuned to produce outputs that appeal to a broad audience. Using them to generate images aligned with individual users relies on iterative manual prompt engineering by the user which is inefficient and undesirable. We propose to personalize the image generation process by first capturing the generic preferences of the user in a one-time process by inviting them to comment on a small selection of images, explaining why they like or dislike each. Based on these comments, we infer a user's structured liked and disliked visual attributes, i.e., their visual preference, using a large language model. These attributes are used to guide a text-to-image model toward producing images that are tuned towards the individual user's visual preference. Through a series of user studies and large language model guided evaluations, we demonstrate that the proposed method results in generations that are well aligned with individual users' visual preferences.