From Unlearning to UNBRANDING: A Benchmark for Trademark-Safe Text-to-Image Generation
This addresses a critical issue for brand owners and AI developers by highlighting a distinct, practically relevant problem in trademark-safe image generation, though it is incremental in focusing on fine-grained brand removal beyond prior work.
The paper tackles the problem of unauthorized reproduction of trademarked content in text-to-image generation by introducing unbranding, a task for removing both explicit trademarks and subtle structural brand features while preserving semantic coherence, and constructs a benchmark dataset with a novel VLM-based evaluation metric to address this.
The rapid progress of text-to-image diffusion models raises significant concerns regarding the unauthorized reproduction of trademarked content. While prior work targets general concepts (e.g., styles, celebrities), it fails to address specific brand identifiers. Crucially, we note that brand recognition is multi-dimensional, extending beyond explicit logos to encompass distinctive structural features (e.g., a car's front grille). To tackle this, we introduce unbranding, a novel task for the fine-grained removal of both trademarks and subtle structural brand features, while preserving semantic coherence. To facilitate research, we construct a comprehensive benchmark dataset. Recognizing that existing brand detectors are limited to logos and fail to capture abstract trade dress (e.g., the shape of a Coca-Cola bottle), we introduce a novel evaluation metric based on Vision Language Models (VLMs). This VLM-based metric uses a question-answering framework to probe images for both explicit logos and implicit, holistic brand characteristics. Furthermore, we observe that as model fidelity increases, with newer systems (SDXL, FLUX) synthesizing brand identifiers more readily than older models (Stable Diffusion), the urgency of the unbranding challenge is starkly highlighted. Our results, validated by our VLM metric, confirm unbranding is a distinct, practically relevant problem requiring specialized techniques. Project Page: https://gmum.github.io/UNBRANDING/.