Microsoft develops garment recognition technology

Tech giant Microsoft has developed new technology that can identify a garment from a smartphone picture, even against a busy background. This promises to be useful for the garment industry.

Many fashion houses do not have traditional seasons and instead continually design new clothes, often influenced by social media. Warehouses use a garment rails supplier to provide well-designed and robust rails that can store hundreds of different garment styles. The difficulty then is identifying each garment hung on the rails, particularly with a high turnover of new stock. There are electronic tag solutions, but these can be expensive.

The Microsoft system is a way to match catalogue photographs with stock to make sure that the correct items are stored or picked.

The developers of the system, CY Yam, Patty Ryan and Elena Terenzi, said on a Microsoft blog:

“Our aim was to work with this retailer to design an algorithm capable of identifying whether a newly arrived item was in stock using only a mobile phone image of the new item as a reference. If the retailer’s staff could snap a photo of a new arrival and use our solution to search their catalogue of studio images for matches, we could eliminate the cost of errors and wasted time.”

A major hurdle that the developers faced was separating the clothing item from the background without having to manually isolate the foreground. They managed this so that identifying garments is both quick and easy. It is hoped that this technology could help retailers manage and control their inventory more efficiently.

You may also interested in: