Roorkee - Indian Institute of Technology, Roorkee, has achieved a pioneering milestone in artificial intelligence and creative design with the development of the FLORA dataset and the NeRA (Nonlinear Expressive Representation Adapter) - two groundbreaking innovations that redefine how AI understands, interprets, and generates fashion designs. 

The research, titled “Dressing the Imagination: A Dataset for AI-Powered Translation of Text into Fashion Outfits and a Novel NeRA Adapter for Enhanced Feature Adaptation”, showcases India’s first framework that enables AI models to generate detailed, high-fidelity fashion sketches directly from textual descriptions. Unlike existing fashion datasets that rely mainly on photographs or catalog images, FLORA connects professional fashion sketches with rich textual design language, including silhouettes, materials, textures, and overall style.

The advancement allows AI to understand complex fashion descriptions such as specific cuts, drapes colours and accessories rather than simple keywords. Alongside FLORA, the NeRA adapter in enhances AI learning efficiency, enabling faster and more accurate interpretation and intricate design details.The open source release of both tools is expected to empower designers and researchers worldwide to collaborate at the intersection of art, technology and sustainability. 

The research is led by Sparsh Mittal, who highlighted that the work brings AI and design together to make creativity easier, faster, and more sustainable. He noted that modern fashion relies heavily on subtle details that many AI systems fail to grasp, and FLORA with NeRA bridges this critical gap by teaching AI the language designers actually use.

Congratulating the research team, Kamal Kishore Pant, Director of IIT Roorkee, stated that the achievement reflects the institute’s commitment to cutting-edge interdisciplinary research and demonstrates how AI can strengthen India’s creative economy while enhancing global leadership in design and technology.

The innovations aligns with national initiatives such as Digital India, Make in India and the National AI Mission, promoting indigenous technologies that support sustainable design solutions. By integrating AI into fashion design, the research shortens design cycle, democratizes creative tools and open new  possibilities in personalized fashion, virtual retail, gaming, metaverse, application and sustainable manufacturing. 

The study also supports the United Nations Sustainable Development Goals, particularly SDG 9 (Industry, Innovation and Infrastructure) and SDG 12 (Responsible, Consumption and Production) reinforcing India's growing leadership in socially and environmentally responsible AI innovation. 

The research paper is co-authored by Gayatri Deshmukh (Independent Researcher), Somsuhbra De (IIT Madras), Chirag Sehgal (Delhi Technological University), Jishu Sen Gupta (IIT BHU), and Prof. Sparsh Mittal (IIT Roorkee). With this breakthrough, IIT Roorkee further establishes itself as a frontrunner in generative AI research with far-reaching societal and creative impact.

.

Click Here for More Latest News