Personalized Image Generation with Reasoning and Reflection
cs.CV, cs.AI
Submitted: 2026-09-30
Updated: 2026-09-30
License: http://creativecommons.org/licenses/by/4.0/
The gist: Personalized image generation has remained narrowly focused on conditional synthesis from curated visual exemplars, rather than capturing who a user is.
Terminology
Abstract
Personalized image generation has remained narrowly focused on conditional synthesis from curated visual exemplars, rather than capturing who a user is. In practice, however, a user's personal context is much richer, comprising reviews, posts, images, captions, and metadata accumulated over time. A truly personalized generator should leverage this history to produce images aligned with the user's lifestyle and aesthetic preferences. To this end, we introduce the first unified benchmark for personalized image generation from user histories. The benchmark comprises two complementary tasks and a multi-axis evaluation protocol that assesses target fidelity, visual quality, user distinguishability, semantic alignment with the user's history, and task-specific utility. Grounded in real-world e-commerce and social media settings, the benchmark includes: (1) Personalized Scene Generation, which places a given object in a scene that reflects a user's preferences and lifestyle, motivated by personalized product presentation; and (2) Personalized Creative Generation, which generates a novel image on a specified topic that is faithful to a user's aesthetic and visual identity, motivated by social media content creation. We further propose PEARL, which couples a multimodal reasoner with a frozen image generator in an interleaved reason-reflect loop optimized with differential data reward. Across both tasks, PEARL outperforms strong baselines, achieving an average improvement of 15% across personalization metrics.
Sources
- Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond
- Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
- Personalized Image Editing in Text-to-Image Diffusion Models via Collaborative Direct Preference Optimization
- Bridging Language and Items for Retrieval and Recommendation: Benchmarking LLMs as Semantic Encoders
- LoRA: Low-Rank Adaptation of Large Language Models
- Interleaving Reasoning for Better Text-to-Image Generation
- Unified Language-Vision Pretraining in LLM with Dynamic Discrete Visual Tokenization
- LongLaMP: A Benchmark for Personalized Long-form Text Generation
- Personalized Language Modeling from Personalized Human Feedback
- Visual Instruction Tuning
- Deep Learning Recommendation Model for Personalization and Recommendation Systems
- Reasoning-Based Personalized Generation for Users with Sparse Data
- Direct Preference Optimization: Your Language Model is Secretly a Reward Model
- DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generation
- LaMP: When Large Language Models Meet Personalization
- LAION-5B: An open large-scale dataset for training next generation image-text models
- Personalized Image Generation with Deep Generative Models: A Decade Survey
- Controllable Textual Inversion for Personalized Text-to-Image Generation
Related papers
- Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
- AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
- Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
- MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
- TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
- A Survey on Efficient Vision-Language-Action Models