Verify Claimed Text-to-Image Models via Boundary-Aware Prompt Optimization
Authors: Zidong Zhao, Yihao Huang, Qing Guo, Tianlin Li, Anran Li, Kailong Wang, Jin Song Dong, Geguang Pu
Organizations: Zhejiang University, Hangzhou, China · East China Normal University, Shanghai, China · Nankai University, Tianjin, China · Beihang University, Beijing, China · University of Science and Technology of China, Hefei, China · Huazhong University of Science and Technology, Wuhan, China · National University of Singapore, Singapore, Singapore
As Text-to-Image (T2I) generation becomes widespread, third-party platforms increasingly integrate multiple model APIs for convenient image creation. However, false claims of using official models can mislead users and harm model owners' reputations, making model verification essential to confirm whether an API's underlying model matches its claim. Existing methods address this by using verification prompts generated by official model owners, but the generation relies on multiple reference models for optimization, leading to high computational cost and sensitivity to model selection. To address this problem, we propose a reference-free T2I model verification method called Boundary-aware Prompt Optimization (BPO). It directly explores the intrinsic characteristics of the target model. The key insight is that although different T2I models produce similar outputs for normal prompts, their semantic boundaries in the embedding space (transition zones between two concepts such as "corgi" and "bagel") are distinct. Prompts near these boundaries generate unstable outputs (e.g., sometimes a corgi and sometimes a bagel) on the target model but remain stable on other models. By identifying such boundary-adjacent prompts, BPO captures model-specific behaviors that serve as reliable verification cues for distinguishing T2I models. Experiments on five T2I models and four baselines demonstrate that BPO achieves superior verification accuracy.
Figures & tables
Fig. 1: API provider may claim an official model while actually serving a cheaper one.
Fig. 2: The process of the model verification.
Fig. 3: Analysis of semantic boundaries via interpolation in different T2I models.
Fig. 4: Overview of the BPO method.
Model
Method
Accuracy
Precision
Recall
F1-Score
Stable Diffusion v1.4
Normal
0.17
0.17
1.00
0.29
Random
0.33
0.20
1.00
0.33
Greedy
0.17
0.17
1.00
0.29
TVN
0.50
0.25
1.00
0.40
BPO (Ours)
1.00
1.00
1.00
1.00
Stable Diffusion v2.1
Normal
0.17
0.17
1.00
0.29
TABLE I: Performance of model verification compared with baselines.
Fig. 5: Comparison of generated images between the target model and other models using the same verification prompt.
Model
Prompt
Accuracy
Precision
Recall
F1-Score
Stable Diffusion v1.4
Ppis
0.40
0.25
1.00
0.40
Padv
0.60
0.33
1.00
0.50
Pv
1.00
1.00
1.00
1.00
Stable Diffusion v2.1
Ppis
0.60
0.33
1.00
0.50
Padv
1.00
1.00
1.00
1.00
Pv
0.80
0.50
1.00
0.67
TABLE II: Ablation performance of each prompt in the three-stage BPO pipeline on model verification.
Fig. 6: Verification performance with varying numbers of generated images in the model verification phase.
Fig. 7: Verification performance with varying suffix lengths.
Method
Time (s)
SDXL
SD2.1
SD1.4
Dreamlike
Openjourney
TVN
366.66
224.72
553.12
256.97
203.87
BPO
323.47
149.88
108.11
121.34
92.60
TABLE III: The average time for TVN and BPO methods to generate one verification prompt on each model.
Text-to-image (T2I) generation models have made significant strides but still struggle with prompt sensitivity: even minor changes in prompt wording can yield inconsistent or inaccurate outputs. To address this challenge, we introduce a closed-loop, test-time prompt refinement framework that requires no additional training of the underlying T2I model, termed TIR. In our approach, each generation step is followed by a refinement step, where a pretrained multimodal large language model (MLLM) analyzes the output image and the user's prompt. The MLLM detects misalignments (e.g., missing objects, incorrect attributes) and produces a refined and physically grounded prompt for the next round of image generation. By iteratively refining the prompt and verifying alignment between the prompt and the image, TIR corrects errors, mirroring the iterative refinement process of human artists. We demonstrate that this closed-loop strategy improves alignment and visual coherence across multiple benchmark datasets, all while maintaining plug-and-play integration with black-box T2I models. Code is available at https://github.com/hafeezkhan909/Test-time-Image-Refinement.
Mohammad Abdul Hafeez Khan, Yash Jain, Siddhartha Bhattacharyya +1
Florida Institute of Technology, USA · Microsoft Research, USA
The rapid advancements of Text-to-Image (T2I) models have ushered in a new phase of AI-generated content, marked by their growing ability to interpret and follow user instructions. However, existing T2I model evaluation benchmarks fall short in limited prompt diversity and complexity, as well as coarse evaluation metrics, making it difficult to evaluate the fine-grained alignment performance between textual instructions and generated images. In this paper, we present TIIF-Bench Text-to-Image Instruction Following Benchmark), aiming to systematically assess T2I models' ability in interpreting and following intricate textual instructions. TIIF-Bench comprises 5,000 prompts organized along multiple dimensions and categorized into three levels of difficulty and complexity. To rigorously evaluate robustness to prompt length, each prompt is provided in both short and long versions with identical core semantics. We further propose a novel Global Normalized Edit Distance (GNED) metric for text rendering and provide aspect-ratio-diverse reference images for each prompt to assess style control. In addition, we collect 100 high-quality designer-level prompts covering diverse scenarios for comprehensive evaluation. To enable scalable and fine-grained evaluation, we explore the best paradigm for leveraging the world knowledge encoded in large Vision-Language Models (VLMs) as automated binary evaluators. Through extensive ablations, we develop a fully reproducible evaluator that provides interpretable reasoning and reliable verification, enabling our benchmark to discern subtle variations in T2I model outputs. Through comprehensive benchmarking of mainstream T2I models on TIIF-Bench, we analyze the strengths and weaknesses of current T2I systems and reveal the limitations of existing evaluation benchmarks. Project Page: https://a113n-w3i.github.io/TIIF_Bench/.
Xinyu Wei, Jinrui Zhang, Zeqing Wang +4
The Hong Kong Polytechnic University · OPPO Research Institute · Tsinghua University
Red-teaming Text-to-Image (T2I) models is essential for safe deployment, yet it remains particularly challenging against implicit adversarial prompts. Unlike explicit adversarial prompts that can be readily identified and blocked, implicit ones are much harder to detect: the prompts appear benign on the text surface yet still lead to inappropriate visual content. To address this, we propose Adversarial Probing for Implicit VulnErabilities (AdvPIE), a multimodal agentic framework to expose implicit vulnerabilities without requiring access to the parameters of target models. AdvPIE adopts a policy agent to generate and refine implicit adversarial prompts based on the feedback from a judge agent. To construct informative feedback, the judge agent provides modality-specific safety evaluation at both global and relative levels across iterations. To effectively leverage the feedback, we propose a novel Cumulative Adversarial Decoding strategy for the policy agent, which dynamically reweights token distributions to favor tokens that lead to more harmful images while preserving sampling diversity. Extensive experiments on standard and safety-aligned T2I models show that AdvPIE1 effectively uncovers implicit vulnerabilities, outperforming various baseline methods.
Chang Ma, Junlin Han, Shuo Chen +3
Columbia University, USA · University of Oxford, UK · LMU Munich, Germany +2