WGSN Unlocks Fashion Trends With AI Image Analysis

Ever wondered how the torrent of images from global fashion weeks is translated into concrete, actionable trends? With five million images to analyze, trend forecasting leader WGSN gives us a behind-the-scenes look at how they break down fashion trends in real time, blending human expertise with powerful AI.

We dive into the process with Cassandra Rosenthal, Senior Digital Asset Manager at WGSN, to explore the data-driven advances in image tagging that help brands access trends faster and more accurately than ever before.

The Scale of a Global Fashion Library

At the heart of WGSN's operation is a colossal digital asset library. "My role at WGSN is digital asset management," explains Rosenthal. "I lead a large global team of fashion, beauty, interiors, material and sports experts who work to categorise data from all our imagery." She compares the work to that of a digital librarian, organizing and evaluating vast quantities of information.

This is no small library. The WGSN Catwalk Library holds over five million images, documenting more than 25 years of fashion history. It covers every single look from the major catwalks in New York, London, Milan, and Paris for both main and pre-season collections, with coverage continuously expanding to more regional shows globally.

The Art and Science of Image Tagging

So, how does an image become data? A global team of fashion experts meticulously reviews every full-length runway photo. Using a proprietary system, they break down each look into 27 different garment and accessory categories. From there, images are cropped, color-tagged, and enriched with highly specific details.

For example, a single dress is analyzed and tagged with incredible granularity:

• Product: Cocktail Dress, Empire/Babydoll, Bubble/Balloon
• Materials: Crinkle/Wrinkle, Shine
• Colour: Neutral
• Embellishment/Trim: Ruffles, Embroidery
• Length: Knee Length
• Sleeves: Short Sleeve, Puff/Gathered Sleeve
• Neckline: Sweetheart Neck
• Print/Pattern: Florals
• Design Details: Backless

AI as a Catalyst for Speed and Accuracy

Recently, WGSN has supercharged this process by incorporating artificial intelligence. "We sped up data delivery to our clients by leveraging AI to tag, crop and colour-tag key categories within minutes of receiving Catwalk data," Rosenthal notes. This initial pass by the AI provides a massive speed advantage.

Crucially, human expertise remains central to the process. WGSN's experts perform a "fast follow," reviewing the AI model's work to amend, refine, and add nuanced information. This human-in-the-loop system ensures the final data meets WGSN's high standards for accuracy.

From Runway Image to Actionable Trend Forecast

This tagged imagery is the first crucial input for WGSN’s trend forecasting. The runway is often the first place to spot emerging trends, color shifts, and changes in category popularity. This data is then analyzed using WGSN’s STEPIC Index, which combines these qualitative insights with quantitative data from e-commerce shelves, search traffic, and social media to forecast trends up to 24 months into the future.

Empowering Brands with Faster Insights

By delivering more data more quickly, WGSN helps clients immediately understand catwalk trends and integrate them into their next season's assortment plans. An AI model that combines catwalk, social, and retail data helps assess a trend's strength over time, empowering brands to make confident decisions on which styles to champion.

Imagine having all the catwalk analysis you need, from live coverage to post-show analytics, in one place. WGSN enables you to quickly identify emerging and evolving trends every season.

Ready to see how it works? Access the future of trend forecasting now.