Harnessing Foundation Models for Image Anonymization

Luca Piano, Pietro Basci, Fabrizio Lamberti, Lia Morra

Abstract: Traditional deep learning pipelines involve multiple intricate steps, from data acquisition to model training, fine-tuning, and deployment. However, recent advancements in foundation models, particularly in text-to-image generation, offer a paradigm shift in addressing tasks without the need for these conventional processes. In this paper, we explore how foundation models can be leveraged to solve tasks, specifically focusing on anonymization, without the requirement for training or fine-tuning. By bypassing traditional pipelines, we demonstrate the efficiency and effectiveness of this approach in achieving anonymization objectives directly from the foundation model's inherent knowledge. Our findings underscore the transformative potential of foundation models in simplifying and accelerating deep learning tasks, paving the way for novel applications in various domains.