AlbumFill: Album-Guided Reasoning and Retrieval for Personalized Image Completion

2026-05-04

AI Summary

Personalized image completion struggles to preserve identity and appearance consistency in the absence of explicit reference images. To address this challenge, this work proposes a training-free framework that, for the first time, integrates album-guided semantic reasoning and retrieval into the task. Specifically, a vision-language model infers semantic cues about the missing region, which are then used to automatically retrieve identity-consistent reference images from the user’s personal photo album. These retrieved references subsequently guide a reference-based inpainting model to achieve high-fidelity reconstruction. Evaluated on a new dataset comprising 54K portrait samples, the proposed method significantly outperforms existing baselines, demonstrating the critical role of identity-consistent reference retrieval in enabling effective personalized image completion.

Abstract

Personalized image completion aims to restore occluded regions in personal photos while preserving identity and appearance. Existing methods either rely on generic inpainting models that often fail to maintain identity consistency, or assume that suitable reference images are explicitly provided. In practice, suitable references are often not explicitly provided, requiring the system to search for identity-consistent images within personal photo collections. We present AlbumFill, a training-free framework that retrieves identity-consistent references from personal albums for personalized completion. Given an occluded image and a personal album, a vision-language model infers missing semantic cues to guide composed image retrieval, and the retrieved references are used by reference-based completion models. To facilitate this task, we introduce a dataset containing 54K human-centric samples with associated album images. Experiments across multiple baselines demonstrate the difficulty of personalized completion and highlight the importance of identity-consistent reference retrieval. Project Page: https://liagm.github.io/AlbumFill/

Problem

• personalized image completion
• identity consistency
• reference retrieval
• image inpainting
• personal photo albums

Innovation

• personalized image completion
• identity-consistent retrieval
• vision-language model
• reference-based inpainting
• album-guided reasoning