Many tech corporations have give you display screen reader software program for visually impaired folks to make it simple for them to make use of social media platforms. Nevertheless, till now there was no expertise out there to translate memes that don’t have various textual content or alt textual content to explain the meme.
Now, researchers from Carnegie Mellon College have invented a technology that identifies memes and provides descriptive alt textual content to make it accessible for visually impaired folks.
Based on Jeff Bigham, affiliate professor in CMU’s Human-Pc Interplay Institute (HCII), “Memes might not appear to be crucial drawback, however an important a part of accessibility will not be selecting for folks what deserves their consideration.”
Often, it’s not attainable for meme creators so as to add alt textual content on social media the place a lot of the memes are discovered. Twitter, nevertheless, permits customers so as to add alt textual content the photographs they’re importing however the characteristic is troublesome to seek out and most customers usually ignore it.
A research performed by CMU researchers confirmed that solely 0.1% of the a million photograph tweets had alt textual content.
The brand new expertise makes use of fundamental pc imaginative and prescient strategies for figuring out the meme template and OCR strategies are deployed for studying the overlaid textual content on memes. Researchers are attempting to standardize a template for describing a meme and the overlaid textual content may be added based on a selected meme.
Nevertheless, changing a meme into accessible textual content for visually impaired continues to be a problem.
As Cole Gleason, a Ph.D. scholar at HCII says, “It relied on the meme if the humor translated. A number of the visuals are extra nuanced”
Researchers have additionally created a platform for changing memes into audio format. The platform permits customers to pull and drop audio information into meme templates. The system can also be able to translating current memes into audio memes to convey the message via sound results.
Gleason says, “One of many causes we tried the audio memes was as a result of we thought alt textual content would kill the joke, however folks nonetheless most popular the textual content as a result of they’re so used to it.”
Now that researchers have developed the expertise, the following step is to combine it into numerous meme generator web sites and social media platforms.
Researchers are fascinated with talking to Twitter for including a brand new characteristic to make it simple for including alt textual content. CMU researchers are additionally engaged on creating a browser extension for Twitter so as to add alt textual content for each picture.
In one other associated undertaking, researchers are discovering a way to combine alt textual content into the metadata uploaded with the picture. It could append the alt textual content with the picture whatever the platform it’s uploaded on.
Additionally Learn: Lenovo’s Leez Single Board Computer Is Here To Rival Raspberry Pi Boards