Google Cloud unveils AI tool to tackle shelf inventory accuracy

Ahead of the NRF Big Show taking place this week in NYC, Alphabet's cloud division announced the development of a new machine vision algorithm capable of accurately taking stock of on-shelf inventory via cameras mounted on ceilings or used by employees (either the human or robotic). The system draws on a database of over a billion product images and can perform at over 90 percent accuracy, according to early testing at the Giant Eagle chain. The new development is being offered as an enhancement to the Google Cloud for Retailers suite, updating the firm's Discovery AI and Recommendation AI services.