CopyCat: Improving Fine-Grained Subject Consistency in Subject-to-Image Models within Seconds

• URL: http://arxiv.org/abs/2608.00674v1
• Date: Sat, 01 Aug 2026 13:49:03 GMT
• Title: CopyCat: Improving Fine-Grained Subject Consistency in Subject-to-Image Models within Seconds
• Abstract summary: CopyCat is a lightweight model-refinement framework that improves fine-grained subject consistency within only a few seconds.<n>Experiments on DreamBench and XVerseBench demonstrate consistent improvements in fine-grained subject consistency across representative subject-to-image models.
• Score: 52.84219088497834
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Recent subject-to-image models have achieved impressive progress in personalized image generation, yet they still struggle to preserve fine-grained subject-specific details. A major reason is the lack of high-quality fine-grained identity supervision: real paired data are expensive to collect, while synthesized training pairs often preserve only coarse subject appearance and fail to capture subtle subject-specific details. In this work, we propose CopyCat, a lightweight model-refinement framework that improves fine-grained subject consistency within only a few seconds. CopyCat performs a one-time refinement of a pretrained subject-to-image model by attaching a lightweight Fine-grained Consistency LoRA (FCLoRA) and optimizing it using a single proxy image, which is used as both the conditioning image and the reconstruction target. This exact self-reconstruction objective substantially simplifies the optimization task, enabling effective fine-grained refinement within only a few seconds. The refinement is performed only once; the resulting model can be directly applied to diverse unseen reference subjects and prompts without further subject-specific optimization. We further revisit subject-to-image LoRA training in double-stream diffusion transformers and find that adapting only the visual stream consistently improves subject consistency. Extensive experiments on DreamBench and XVerseBench demonstrate consistent improvements in fine-grained subject consistency across representative subject-to-image models under both single- and multi-subject settings.

Related papers

• Universal Image Restoration via Internalized Chain-of-Thought Reasoning [52.763760468548504]
  Image restoration seeks to recover high-quality images from degraded inputs but becomes highly ill-posed under complex, mixed degradations.<n>Recent works adopt Chain-of-Thought (CoT) reasoning for multi-round restoration using specialized modules.<n>We introduce CoTIR, a universal image restoration framework that internalizes CoT reasoning within a single model.
  arXiv  Detail & Related papers  (2026-06-16T06:01:36Z)
• Edit2Restore:Few-Shot Image Restoration via Parameter-Efficient Adaptation of Pre-trained Editing Models [4.573600918393017]
  We show that powerful pre-trained text-conditioned image editing models can be efficiently adapted for multiple restoration tasks.<n>Our approach fine-tunes LoRA adapters on FLUX.1 Kontext, a state-of-the-art 12B parameter flow matching model for image-to-image translation.
  arXiv  Detail & Related papers  (2026-01-06T19:56:16Z)
• From Wardrobe to Canvas: Wardrobe Polyptych LoRA for Part-level Controllable Human Image Generation [44.46447676191666]
  We present Wardrobe Polyptych LoRA, a controllable part-level controllable model for personalized human image generation.<n>By training only LoRA layers, our method removes the computational burden at inference while ensuring high-fidelity synthesis of unseen subjects.<n>Our approach significantly outperforms existing techniques in fidelity and consistency, enabling realistic and identity-preserving full-body synthesis.
  arXiv  Detail & Related papers  (2025-07-14T12:34:25Z)
• Single Image Iterative Subject-driven Generation and Editing [40.285860652338506]
  We present SISO, a training-free approach to personalize image generation and editing from a single image without training.<n>SISO iteratively generates images and optimize the model based on loss of similarity with the given subject image.<n>We demonstrate significant improvements over existing methods in image quality, subject fidelity, and background preservation.
  arXiv  Detail & Related papers  (2025-03-20T10:45:04Z)
• EZIGen: Enhancing zero-shot personalized image generation with precise subject encoding and decoupled guidance [20.430259028981094]
  Zero-shot personalized image generation models aim to produce images that align with both a given text prompt and subject image.<n>Existing methods often struggle to capture fine-grained subject details and frequently prioritize one form of guidance over the other.<n>We introduce a new approach, EZIGen, that employs two main components: leveraging a fixed pre-trained Diffusion UNet itself as subject encoder.
  arXiv  Detail & Related papers  (2024-09-12T14:44:45Z)
• Is Synthetic Image Useful for Transfer Learning? An Investigation into Data Generation, Volume, and Utilization [62.157627519792946]
  We introduce a novel framework called bridged transfer, which initially employs synthetic images for fine-tuning a pre-trained model to improve its transferability. We propose dataset style inversion strategy to improve the stylistic alignment between synthetic and real images. Our proposed methods are evaluated across 10 different datasets and 5 distinct models, demonstrating consistent improvements.
  arXiv  Detail & Related papers  (2024-03-28T22:25:05Z)
• Exposure Bracketing Is All You Need For A High-Quality Image [50.822601495422916]
  Multi-exposure images are complementary in denoising, deblurring, high dynamic range imaging, and super-resolution.<n>We propose to utilize exposure bracketing photography to get a high-quality image by combining these tasks in this work.<n>In particular, a temporally modulated recurrent network (TMRNet) and self-supervised adaptation method are proposed.
  arXiv  Detail & Related papers  (2024-01-01T14:14:35Z)
• Separate-and-Enhance: Compositional Finetuning for Text2Image Diffusion Models [58.46926334842161]
  This work illuminates the fundamental reasons for such misalignment, pinpointing issues related to low attention activation scores and mask overlaps. We propose two novel objectives, the Separate loss and the Enhance loss, that reduce object mask overlaps and maximize attention scores. Our method diverges from conventional test-time-adaptation techniques, focusing on finetuning critical parameters, which enhances scalability and generalizability.
  arXiv  Detail & Related papers  (2023-12-10T22:07:42Z)
• Diversified in-domain synthesis with efficient fine-tuning for few-shot classification [64.86872227580866]
  Few-shot image classification aims to learn an image classifier using only a small set of labeled examples per class. We propose DISEF, a novel approach which addresses the generalization challenge in few-shot learning using synthetic data. We validate our method in ten different benchmarks, consistently outperforming baselines and establishing a new state-of-the-art for few-shot classification.
  arXiv  Detail & Related papers  (2023-12-05T17:18:09Z)
• ReGeneration Learning of Diffusion Models with Rich Prompts for Zero-Shot Image Translation [8.803251014279502]
  Large-scale text-to-image models have demonstrated amazing ability to synthesize diverse and high-fidelity images. Current models can impose significant changes to the original image content during the editing process. We propose ReGeneration learning in an image-to-image Diffusion model (ReDiffuser)
  arXiv  Detail & Related papers  (2023-05-08T12:08:12Z)
• Robust Single Image Dehazing Based on Consistent and Contrast-Assisted Reconstruction [95.5735805072852]
  We propose a novel density-variational learning framework to improve the robustness of the image dehzing model. Specifically, the dehazing network is optimized under the consistency-regularized framework. Our method significantly surpasses the state-of-the-art approaches.
  arXiv  Detail & Related papers  (2022-03-29T08:11:04Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.