Persistent Identity Preservation in Generative Image Models: A Benchmark and Evaluation System
2026-09-03 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors studied how well current AI image generators keep the same identity when creating or editing pictures of a person or object. They found that even if images look good and follow instructions, the identity can still change a lot, especially with multiple edits, small subjects, or complex scenes. They tested different ways to store identity information and showed that having a persistent identity representation helps keep things consistent across generations and edits. This means that identity needs special handling separate from the general image generation process.
Generative image modelsIdentity preservationSubject-driven generationImage editingLoRAPHOTA IDENTITYMulti-subject compositionIterative editsFoundation models
Authors
Mengwei Ren, Xuaner Zhang, Zhihao Xia
Abstract
Generative image models can now produce high-quality images, follow complex instructions, and support precise edits, but they still struggle to preserve who or what is being depicted. When generating or editing images of a specific subject, identity may drift as the pose, expression, appearance, viewpoint, or surrounding scene changes. Existing subject-driven methods make fundamentally different choices about where identity is represented: through the input context (GPT-Image-2, NB2), as trainable subject-specific model parameters (LoRA), or as a persistent identity layer (PHOTA IDENTITY) reusable across generations and edits. We systematically benchmark these paradigms across subject-driven generation, editing, restoration, and multi-subject settings, with tasks designed to increasingly stress identity preservation. Our results show that identity preservation remains a distinct limitation of current generative foundation models: strong image quality and instruction following do not necessarily imply strong identity fidelity, and identity degradation becomes more pronounced under iterative edits, small subject scales, severe image degradation, and multi-subject composition. Persistent identity substantially reduces this degradation across generation, editing, and restoration, consistently improving identity preservation when applied to different foundation models while maintaining comparable instruction adherence and perceptual image quality. These results suggest that identity does not simply emerge from increasingly capable generative models, but can instead be represented as persistent subject knowledge that is composed independently with the underlying generative model.