Clothes Design utilizing Generative AI

Abstract

The dynamic domain of text-to-image generation has witnessed remarkable advancements in recent years, particularly in applications like fashion design and artistic creation, where AI algorithms have demonstrated their capabilities. In addressing the intrinsic challenge of acquiring well-labeled datasets for training, we embraced a framework outlined in [7]. This framework hinges on the collaboration of two pivotal models: the Prior Model and the Decoder. Through careful curation of publicly available datasets, we meticulously prepared datasets for both pre-training and fine-tuning phases, with a deliberate choice of the MS-COCO dataset owing to its suitability and content integrity. Our efforts lay in the development of a comprehensive pipeline for text-to-image generation, complemented by a demo for deploying a model API. This demonstrated the potential of our approach in establishing a user-friendly service for designing Print On Demand images. The efforts not only showcase the current state-of-the-art in AI-driven image generation but also position our framework as a practical and accessible solution for users seeking personalized and creative image design experiences.