Speak contents of images for the visually impaired
Hi friends, Machina here with a new challenge perfect for an individual developer to try (and blog about later) or as part of a hackathon.
We believe in helping people where possible, and we think machine learning is a great tool to do it.
In the US alone, more than 3.4 million (3%) Americans aged 40 years and older are either legally blind or are visually impaired.
One unique challenge to these individuals is having images on websites be inaccessible. There is supposed to be information in the <img alt> attribute in the HTML, but let’s be honest, it isn’t always there, and it doesn’t address issues where users are uploading content to an aggregation site.
So how can we help using machine learning tools like Tagbox?
Your challenge is to build an app, widget, website or otherwise that speaks the content of an image.
We suggest first trying Tagbox, which comes pre-trained with thousands of labels. If you give it a photo, it will return tags that describe what it thinks the photo is all about.
Perhaps the readers you are targeting with your app are into travel, so teaching Tagbox a few landmarks might improve the experience.
Whatever you come up with, please let us know about it and share it with the community. We’d love to highlight your work!
