AI technologies are revolutionizing the fashion industry, but available datasets are generally western-centric. InFashAI creates datasets and AI models that are truly representative of the fashion world's diversity — starting with African fashion.
AI models trained on under-diversified datasets have limited generalization capabilities and may not perform well on African fashion styles — limiting the adoption of AI within the African fashion industry. Ai4Innov initiated InFashAI to build a large volume of data on African fashion, making it the backbone for AI tools adapted to African fashion.
Watch our AI-powered captioning tool automatically describing African fashion images.
We curated two versions of the InFashAI dataset by gathering scattered African fashion data accessible online using web scraping methods.
Each image comes with its description — ready for fashion image captioning and realistic fashion image generation from text.
From detection to generation, InFashAI brings African fashion into the AI era.
Automatically detects clothing, patterns, and accessories in fashion images.
Generates precise descriptions in multiple languages for global reach.
Showcases African creations on digital platforms worldwide.
Supports realistic fashion image generation from text descriptions.
Built with state-of-the-art research and powerful infrastructure.
Deep learning models for neural image captioning tailored to fashion.
Large-scale automated data collection from diverse online sources.
Cloud infrastructure provided by Oracle for Research.
Industry collaboration with Africa's leading fashion marketplace.
Explore our datasets on GitHub and read the published research.
InFashAI is made possible thanks to the generous support of our partners.
