TrademarkVision is offering deep-learning technology that can help companies determine whether their logos have any infringement issues—and soon may be able to perform the same service for product designs, according to Fastco Design.
The San Francisco-based company uses a visual search engine that allows companies to upload their logo into a window on the system. The system then maps the shape of the image and uses that information to sift through trademark databases throughout the world to see if any of the millions of logos is a match. Corporate customers can use the system both to determine if their logos infringe on others and to learn if others companies’ emblems too closely resemble their own designs.
The system provides a much faster and more accurate solution to trademark search than the conventional approach of manually combing through the multitude of global databases using imprecise methods, such as Vienna codes or keyword searches, TrademarkVision says.
The company is employing a combination of various image-matching algorithms along with a semantic understanding layer that identifies objects in a logo image.
TrademarkVision is experimentally applying this technology to the industrial design of the product. The feature, which is in beta, allows customers to load multiple views of a 3-D product into the system. TrademarkVision’s technology then can search databases to see if it closely resembles an existing design.
Tyler Schulze is vice president, strategy & development at Veritone. He serves as general manager for developer partnerships, cognitive engine ecosystem, and media ingestion for the Veritone platform. Learn more about our platform and join the Veritone developer ecosystem today.