Instagram is using object recognition tech to describe photos for visually impaired users

[NelsonG]

Instagram is a visual service; it’s a feed of photos and videos and memes that users take in, but the company is rethinking how to optimize a visual service for users with visual impairments via a couple of new features.

The accessibility-focused update rolling out today offers two options to give visually impaired users alternative text descriptions of what’s happening in posts. One leverages user text input while the other uses Facebook’s object recognition tech to do the heavy lifting.

The descriptions will surface for users utilizing screen readers. The user-input alternative text option will pop up in advanced settings when a user is posting a photo. It’s definitely a little out of the way, but it exists now, so that’s something. While there’s a level of depth that can only come from a human-written visual description, the company’s use of its object recognition software is obviously going to be carrying out most of the text descriptions on items in the Feed, Explore and Profile sections.

It’s an interesting proposition for a service that has made a name for itself through being almost entirely visual, keeping text and external navigations at bay while promoting a feed that’s all about visual absorption. At this point Instagram is far from fringe, and neither is the segment of the population that has a visual impairment — 285 million according to the blog post announcing the feature.

View the full article