WithEveryone: Unified Planning and Identity Grounding for Group Image Generation
2026-08-20 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors present WithEveryone, a system to create images showing up to ten specific people while keeping their identities clear. It works by planning where each person should be in the image and making sure the faces match the intended individuals directly, instead of relying on less reliable face matching methods. Compared to previous approaches, WithEveryone improves how well the generated faces resemble the target people and reduces errors like copying faces incorrectly. This helps create group images that better respect who is who without repeating the same face by mistake.
identity-preserving image generationface recognitionimage synthesislayout planningtoken embeddingface similaritylatent representationgroup image generationvisual conditioningidentity-layout grounding
Authors
Hengyuan Xu, Qixun Wang, Yiji Cheng, Miles Yang, Zhao Zhong, Wei Cheng, Xingjun Ma, Yu-gang Jiang
Abstract
Identity-preserving image generation becomes increasingly unreliable when a scene must contain many specified people. Beyond retaining each identity, the model must bind every reference to a distinct person and location, while training-time identity losses must establish correspondence among several noisy predicted faces. We introduce WithEveryone, a unified framework for generating group images up to ten reference identities. WithEveryone injects each selected identity as an addressed token, predicts a structured identity--layout plan, and renders the plan as a visual condition. Its key objective, Layout-Grounded ID Loss, uses annotated face regions to supervise the intended identities directly, avoiding unstable embedding-based face matching; ID Representation Forcing additionally trains a prediction for each identity before image synthesis. On an identity-disjoint benchmark, WithEveryone achieves the highest target-context identity similarity, improving face similarity from 0.462 for GPT-Image-2 to 0.499, while reducing copy-paste artifacts from 0.169 to 0.055. It further covers 97.3\% of the requested identities with a duplicate rate of only 2.8\%. These results show that explicit identity--layout grounding enables identity-preserving generation to scale to larger groups without relying on direct reference-face copying.