UGC Platform Chute Can Now See

Chute’s visualization of its new machine vision

Chute is adding eyes to its platform for user-generated content (UGC).

Previously, the platform allowed brands to search hashtags, keywords or comments attached to appropriate images created by social media users — that is, searching by text. Now, the company has developed and is today launching its IRIS image recognition technology for detecting a logo, object or scene in an image.

CEO Ranvir Gujral told me that, while machine vision companies like Ditto Labs, Clarifai, or gazeMetrix (recently acquired by social analytics firm Sysomos) can offer similar vision technology, Chute is the only one that combines machine vision with social marketing tools.

In addition to IRIS, the company is also announcing today a social posting tool that lets marketers send or schedule images for social accounts or conversations. Previously, the platform required users to download the located images and then post them manually.

Additionally, a new Discover feature enables searches that combine text and vision or spot trends, and it allows image rights to be obtained and managed within the same interface.

A screen in Chute Discover, employing IRIS vision technology.

Gujral said that the new vision technology can determine if an image contains such objects as a bottle, a cup or a face. It can’t yet conduct facial recognition so you could find all the images with, say, Kim Kardashian’s face in them, but it does know if some face is there. It can also find such scenes as a sunset or a visit to a beach, and it can locate photos that match existing images, like another picnic scene. Multiple search criteria can also be used, such as photos of cars in picnic shots.

The machine vision currently searches a “firehose” feed of millions of images that users have posted to Instagram and Twitter since that brand’s feed was set up, or of images relating to specific conversations. The company said it will add Facebook by December, and eventually will enable searching of imagery going back months.

Gujral said that one client, Honda, previously employed manual searching to find UGC containing images of their cars, as those kinds of photos create about three times as much engagement as non-car images, in the form of likes or comments about its brand. Now, that searching can be conducted in real-time with vision and text search.

Other Chute clients include NBCUniversal, Condé Nast, the National Basketball Association, the United Nations, the New York Times and Ford Motor Co.