The Consistency Critic: Correcting Inconsistencies in Generated Images via Reference-Guided Attentive Alignment

TL;DR

This paper introduces ImageCritic, a reference-guided post-editing method that detects and corrects inconsistencies in generated images, significantly improving detail accuracy in complex scenarios.

Contribution

The paper proposes a novel attention alignment loss and detail encoder for correcting inconsistencies in generated images using reference-guided post-editing.

Findings

Effective correction of detail inconsistencies in generated images

Significant improvements over existing methods in various scenarios

Automatic multi-round local editing enhances image quality

Abstract

Previous works have explored various customized generation tasks given a reference image, but they still face limitations in generating consistent fine-grained details. In this paper, our aim is to solve the inconsistency problem of generated images by applying a reference-guided post-editing approach and present our ImageCritic. We first construct a dataset of reference-degraded-target triplets obtained via VLM-based selection and explicit degradation, which effectively simulates the common inaccuracies or inconsistencies observed in existing generation models. Furthermore, building on a thorough examination of the model's attention mechanisms and intrinsic representations, we accordingly devise an attention alignment loss and a detail encoder to precisely rectify inconsistencies. ImageCritic can be integrated into an agent framework to automatically detect inconsistencies and correct them with multi-round and local editing in complex scenarios. Extensive experiments demonstrate that ImageCritic can effectively resolve detail-related issues in various customized generation scenarios, providing significant improvements over existing methods.