Towards Training-Free Scene Text Editing

Yubo Li; Xugong Qin; Peng Zhang; Hailun Lin; Gangyan Zeng; Kexin Zhang

cs.CVMar 25, 2026

Towards Training-Free Scene Text Editing

Authors: Yubo Li, Xugong Qin, Peng Zhang, Hailun Lin, Gangyan Zeng, Kexin Zhang

Organizations: Institute of Information Engineering, Chinese Academy of Sciences · State Key Laboratory of Cyberspace Security Defense · School of Cyber Security, University of Chinese Academy of Sciences · School of Cyber Science and Engineering, Nanjing University of Science and Technology

Abstract

Scene text editing seeks to modify textual content in natural images while maintaining visual realism and semantic consistency. Existing methods often require task-specific training or paired data, limiting their scalability and adaptability. In this paper, we propose TextFlow, a training-free scene text editing framework that integrates the strengths of Attention Boost (AttnBoost) and Flow Manifold Steering (FMS) to enable flexible, high-fidelity text manipulation without additional training. Specifically, FMS preserves the structural and style consistency by modeling the visual flow of characters and background regions, while AttnBoost enhances the rendering of textual content through attention-based guidance. By jointly leveraging these complementary modules, our approach performs end-to-end text editing through semantic alignment and spatial refinement in a plug-and-play manner. Extensive experiments demonstrate that our framework achieves visual quality and text accuracy comparable to or superior to those of training-based counterparts, generalizing well across diverse scenes and languages. This study advances scene text editing toward a more efficient, generalizable, and training-free paradigm. Code is available at https://github.com/lyb18758/TextFlow

Figures & tables

Figure 1: Comparison of the pipelines between training-based and training-free methods for scene text editing. Training-based methods require large-scale, high-quality paired data that require high computing resources. The training-free method mostly focuses on the attention map for general objects, but ignores the text accuracy and style consistency.

Figure 2: The overall framework of TextFlow. In the first phase, the source image is encoded into latent representations $\mathbf{z}_{t}$ and $\mathbf{z}_{src}$ via the VAE encoder, which are subsequently processed by the FMS module to generate concatenated representations $\mathbf{z}^{src,cat}_{t}$ and $\mathbf{z}^{tar,cat}_{t}$ . These representations, along with their corresponding text embeddings $\mathbf{e}^{src}_{p}$ and $\mathbf{e}^{tar}_{p}$ , are fed into parallel DiT blocks to compute the velocity field differential $\Delta_{V}$ , ultimately producing the edited latent representation $\mathbf{z}_{edit}$ ; In the second phase, $\mathbf{z}_{edit}$ and the target embedding $\mathbf{e}^{tar}_{p}$ are processed by the AttnBoost DiT (AB-DiT), where concatenation and self-attention operations generate refined text-to-image attention maps that enhance textual rendering accuracy through spatial-aware amplification.

Figure 3: Illustration of the proposed FMS Model. The latent representations $\mathbf{z}_{t}$ and $\mathbf{z}_{src}$ are processed with random noise $\epsilon$ through linear interpolation and vector arithmetic operations to maintain style consistency.

Methods	ScenePair
Methods	SSIM ( $\times 10^{-2}$ ) $\uparrow$	PSNR $\uparrow$	MSE ( $\times 10^{-2}$ ) $\downarrow$	FID $\downarrow$	ACC (%) $\uparrow$	NED $\uparrow$
DiffSTE [ 17 ]	22.76	12.26	7.34	180.15	71.11	0.907
TextDiffuser [ 5 ]	26.99	13.93	5.70	56.67	51.48	0.719
AnyText [ 41 ]	30.73	13.66	6.05	51.44	51.12	0.734
TextFlux [ 50 ]	86.57	17.96	1.83	54.64	80.40	0.911
Flux-fill [ 22 ]	82.73	17.10	2.99	107.83	13.74	0.306

Table 1: Performance comparison of different methods on the ScenePair dataset.

Figure 4: Qualitative Analysis. The compared methods include both training-based STE approaches like DiffSTE [ 17 ] , AnyText [ 41 ] , TextFlux [ 50 ] and recent training-free editing techniques FlowEdit [ 20 ] . We also include the powerful foundational model Flux-Kontext [ 21 ] (F-Kontext), for a more extensive comparison.

Figure 5: Qualitative comparison among different DiT-based methods on a full-size image.

FMS Module	SSIM $\uparrow$	PSNR $\uparrow$	MSE $\downarrow$	FID $\downarrow$
FlowEdit [ 20 ]	87.60	20.89	1.16	25.41
Ours w/o FMS	87.09	20.47	1.35	16.69
Ours w FMS	89.04	22.42	0.97	13.52

Table 2: Ablation of FMS modules for image quality.

AttnBoost Module	ScenePair		ScenePair (Random)
AttnBoost Module	ACC(%) $\uparrow$	NED $\uparrow$	ACC(%) $\uparrow$	NED $\uparrow$
FLUX-Kontext [ 21 ]	78.72	0.920	76.63	0.916
Ours w/o AttnBoost	20.35	0.420	18.84	0.391
Ours w AttnBoost	79.80	0.931	74.52	0.874

Table 3: Ablation of AttnBoost considering text accuracy.

Steps	SSIM $\uparrow$	PSNR $\uparrow$	MSE $\downarrow$	FID $\downarrow$	ACC(%) $\uparrow$	NED $\uparrow$
24	86.80	20.21	1.43	16.94	77.97	0.925
30	87.12	19.86	1.46	23.1	79.90	0.928
42	88.04	22.21	0.97	52.8	79.40	0.926
50	89.30	22.47	0.91	13.53	79.98	0.914
70	87.01	21.02	0.90	12.83	79.88	0.914

Table 4: Ablation of inference steps on ScenePair.

Figure 6: Qualitative ablation studies validate the effectiveness of FMS in style preservation and demonstrate the significant improvement in text rendering accuracy achieved by AttnBoost.

Scheduler	ACC(%) $\uparrow$	NED $\uparrow$
Ours w Euler	78.73	0.920
Ours w Overshoot [ 16 ]	79.90	0.931

Table 5: Ablation of scheduler on the ScenePair dataset.

Figure 7: Computation graph of the full attention map between $\mathbf{z}_{\text{edit}}$ and $\mathbf{e}^{\text{tar}}_{p}$ . The green-highlighted region in the lower-left corner illustrates the text-to-image attention components ( $A_{\text{t2i}}$ ) extracted from each DoubleStream Transformer block, which are subsequently utilized for scheduler enhancement in the text rendering.

Figure 8: Schematic diagram of the Overshoot scheduler. Given the latent representation $\tilde{Z}_{t}$ at time $t$ , the scheduler first advances the learned ODE trajectory beyond the target step to obtain $\hat{Z}_{o}$ , then applies calibrated noise to return to the corrected state $\tilde{Z}_{s}$ . The injected noise is precisely controlled to ensure $\tilde{Z}_{s}$ conforms to the marginal distribution of $X_{s}$ .

Method	ScenePair
Method	SSIM ( $\times 10^{-2}$ ) $\uparrow$	PSNR $\uparrow$	MSE ( $\times 10^{-2}$ ) $\downarrow$	FID $\downarrow$	ACC (%) $\uparrow$	NED $\uparrow$
TextCtrl [ 54 ]	37.56	14.99	4.47	43.78	84.67	0.936
Flux-Text [ 23 ]	86.45	17.95	1.93	54.84	70.94	0.877
TextFlow (Ours)	89.03	22.47	0.91	13.53	79.98	0.914

Table 6: Performance of more methods on the ScenePair dataset.

Type	Method	ScenePair	TamperScene-2k			AnyText-Bench-en
Type	Method	HE	ACC(%) $\uparrow$	NED $\uparrow$	HE	Sen.acc(%)	NED	HE
Training-Based	TextFlux [ 50 ]	6.9	19.70	0.42	5.4	8.13	0.21	6.2
	Flux-Text [ 23 ]	7.1	18.60	0.42	5.9	38.89	0.65	8.1
	Qwen-image [ 47 ]	8.0	10.75	0.37	7.7	3,74	0.15	7.5
	Flux-Kontext [ 21 ]	7.3	17.79	0.45	7.2	19.53	0.39	7.9
	Longcat-Edit [ 38 ]	8.2	0.65	0.29	4.9	5.66	0.26	8.1

Table 7: Performance of different methods on TamperScene and AnyText-Bench. HE represents Human Evaluation.

Stren.	SSIM $\uparrow$	PSNR $\uparrow$	MSE $\downarrow$	FID $\downarrow$	ACC(%) $\uparrow$	NED $\uparrow$
0.2	84.81	18.19	2.67	23.16	75.20	0.895
0.5	85.14	19.83	2.06	19.21	76.92	0.908
0.7	85.92	20.86	1.46	17.25	79.90	0.928
1.0	89.03	22.47	0.91	13.53	79.98	0.914
2.0	89.30	23.47	0.90	14.20	77.62	0.894
5.0	88.47	21.02	0.90	12.83	76.88	0.872

Table 8: Ablation of

V_{\Delta}

strength on ScenePair.

Figure 9: Ablation study on parameter $c$ . The results indicate that configurations with $c=2$ or $c=3$ at 50 steps achieve optimal performance in terms of both generation accuracy and human evaluation scores.

Figure 10: Visualized heatmaps of models with and without AttnBoost.

Method	Setting	ACC(%) $\uparrow$	NED $\uparrow$
TextFlow-Kontext	+ Overs. + AttnBoost	81.16	0.93
TextFlow-Kontext	+ Overs.	79.99	0.91
TextFlow-Kontext	-	78.72	0.92
TextFLux	+ Overs.	81.24	0.92
TextFLux	-	80.40	0.91

Table 9: Ablation of Overshoot scheduler on ScenePair.

Model	Setting	Time(s) $\downarrow$	Mme(GB) $\downarrow$	ACC(%)
F-Kontext	Base	260.21	40.12	17.79
F-Kontext	Base+TextFlow	483.43	41.57	18.75
Longcat-Edit	Base	136.54	37.24	0.65
Longcat-Edit	Base+TextFlow	252.62	38.67	20.95
FlowEdit	Base(FLUX.1dev)	100.62	32.20	5.56

Table 10: Inference time and memory cost on A6000, accuracy on TamperScene.

Figure 11: Comparative analysis of editing performance with additional models on full-size images.

Figure 12: Additional visual results demonstrating robust performance across challenging scenarios, including artistic typography, small-font text, complex layouts, and special symbols.

Figure 13: Extensive study on varying length, style change, and different prompts.

Figure 14: Limitations in editing accuracy and practical utility within complex scenarios.

Training-Free Image Editing

See all 19 papers

TextSculptor: Training and Benchmarking Scene Text Editing
May 20, 2026Yiheng Lin, Siyu Jiao, Xiaohan Lan +12Synthetic Data Augmentation Text-Guided Image Editing
Beijing Jiaotong University · Bytedance
Self-Prompting Diffusion Transformer for Open-Vocabulary Scene Text Editing via In-Context Learning
May 15, 2026Hongxi Li, Tong Wang, Chengjing Wu +6Diffusion Transformer Text-Guided Image Editing
1MT Lab, Meitu Inc., Beijing, China · School of Computer Science & Technology, Beijing Institute of Technology, Beijing, China
TextWand: A Unified Framework for Scene Text Editing
Jun 4, 2026Shuyu Wang, Zhile Guan, Hongxiu Chen +5Text-Guided Image Editing Text Generation
School of Electronic and Computer Engineering, Peking University, Shenzhen, China.

May 20, 2026J/K move · Enter open · S save

TextSculptor: Training and Benchmarking Scene Text Editing

Yiheng Lin, Siyu Jiao, Xiaohan Lan, Wei Zhou+11

Beijing Jiaotong University · Bytedance

Synthetic Data Augmentation Text-Guided Image Editing Diffusion-Based Image Editing

Open paper AskPDF arXiv

Recent advances in Multimodal Large Language Models (MLLMs) and diffusion-based generative models have substantially improved prompt-driven image editing. However, scene text editing remains challenging, as it requires models to precisely modify textual content while preserving visual realism and non-target regions. Current open-source models still lag behind proprietary systems, largely due to the scarcity of high-quality training data and the lack of standardized benchmarks tailored to text editing. To address these challenges, we present TextSculptor, a comprehensive framework for data construction and evaluation of scene text editing. We first develop an automated data construction pipeline that combines text-aware image synthesis with programmatic text rendering and compositing. Based on this pipeline, we build TextSculpt-Data, a large-scale dataset containing 3.2M training samples, including 1.2M OCR-verified text-to-image samples and 2M paired text editing samples with naturally aligned source-target images and strong background consistency. We further introduce TextSculpt-Bench, a benchmark covering four fundamental text editing tasks: text addition, text replacement, text removal, and hybrid editing. To support reliable evaluation, we design a tailored protocol that measures text accuracy, visual quality, and background preservation through OCR-based text alignment, multimodal judgment, and background-region similarity. Extensive experiments show that TextSculptor improves open-source text editing performance and narrows the gap to proprietary models. The data and benchmark are available at https://github.com/linyiheng123/TextSculptor.

Abstract

Training-Free Image Editing

Explore similar work

TextSculptor: Training and Benchmarking Scene Text Editing

Self-Prompting Diffusion Transformer for Open-Vocabulary Scene Text Editing via In-Context Learning

TextWand: A Unified Framework for Scene Text Editing