Fashion e-commerce is today a multi-billion-dollar market, and it is expected to double by 2028. However, according to industry research, approximately 30% of shipped goods are returned, with the primary reasons being the wrong size or color. In particular, customers perceive colors from pictures and videos on e-commerce websites differently, leading to significant waste from unnecessary packaging, shipping, and clothes destruction.
The new project aims to reduce the return rate in fashion industry e-commerce by 25% by 2030 due to color issues or 6% in total returns. If successful, the new technology could set new industry standards by building new AI-driven color-editing solutions for photos and video for professional use.
The project is a collaboration between DTU Electro and photography software company Capture One, and Aalborg University, and is backed by the Danish Innovation Fund with over €1.8 million.
“We are aiming to develop new AI-based tools for coloring video clips for fashion and e-commerce that will help create higher accuracy and reduce the number of returns because of color, thereby cutting unnecessary waste.” – Bjørn Thorup, Product Manager, Capture One.
The project combines world-class researchers in the fields of AI and photonics with the industry-leading software within color editing, Capture One, which is used today by major fashion retailers like ASOS, H&M, and Boozt for their ecommerce photography. This collaboration provides a unique opportunity to develop new compression formats, HDR (High Dynamic Range) monitor support, and coloring of moving objects to achieve the most natural colors.
"We're incredibly excited about using Capture One's unique color science in an industry that is looking for ways to become more sustainable. Playing our part in creating the future of ecommerce, a future with far less waste and many more satisfied customers, is hugely important to us." - Rafael Orta, CEO, Capture One.
To achieve its goals, the project has identified five specific objectives
- To create a breakthrough software product that can handle the coloring of video which matches colors in images
- Determine how human color perception differs from video to photos
- Develop new AI techniques to track moving objects in video
- Create new video compression techniques for color preservation
- Develop new color rendering techniques for HDR/WCG (Wide Color Gamut) and new high-end monitors.
- Innovation Fund Denmark investment: 14.479.014 DKK
- Total budget: 25.357.348 DKK
- Duration: 3½ years
- Official title: AI-ColorFashion (AICF)
Follow the project here.