Accessible Media powered by AI/ ML- To enhance user experience of specially-abled users

Every internet user spends approximately ¾th his time online consuming media (Images, videos or audio). Apart from entertainment, across industries online businesses use as an integral part of functionality e.g. retail apparel companies presenting size chart to customer or different colors of clothes. So if media is an integral part of business, it is important to assess it for accessibility before deploying. Can specially-abled customers limited in hearing, vision or cognition consume it, is an important question every product owner should ask his team. Quality Media enhances overall customer experience of the application.

There can be some concerns with visual media possibly causing health hazards like seizures. A lot has been researched and written on causes of epilepsy caused by images and videos and their nature. Photo sensitivity can be a major cause of epilepsy owing to certain images and videos qualities. (Read here)

Additionally there is ever increasing regulatory emphasis on accessibility (read about litigation risk).

Above said problems become very serious when potential numbers of affected users from business point of view are considered. Approximately 5% of world population have some kind of color blindness and there are currently 3% epilepsy patients who are photosensitive.

A quality media will tick all the aspects of Visual, Auditory and Cognition accessibility. Some of these aspects are covered by leading accessibility guidelines WCAG, ADA etc. but not all. Business should look at these concerns to increase the customer experience and ensure equal treatment of its customers rather than just being compliant. Companies can stand out and win loyalty of a potentially untapped customer base.

AI can help automate the media assessment for accessibility and to some extent fix the problems too. This is all possible due to the advancement of deep learning and Digital Signal processing. Capability to ascertain specific features of the image and audio signals has provided possibilities to perform automated assessment for example: brightness, color contrast, flickering, flashing of the image and SNR, frequency range etc. of audio signals. Additionally AI can simulate a specially abled user with limited visual, auditory or cognitive capabilities to assess the usage. Automated periodic assessment of application to make sure it is compliant to avoid legal issues.

Some example usage to assess and improve the application accessibility with AI/ Ml solutions;

1.      Provide, assess, and update a caption or description of an image for users who cannot see the picture properly.

2.      Ensure there are no flashes, flickers or anything that can cause a seizure

3.      People with atypical color perception (color blindness) may not be able to discriminate between different colors, or may miss key information when coded with color only. Automated color contrast validation and Make slight readjustments of graphical elements like fonts, colors, and even spacing for people with visual impairments.

4.      Since hearing loss is typically frequency-dependent and the user may have usable hearing in some bands, yet none at all in others. AI can make it possible to apply more sophisticated audio processing such as pre-emphasis filters, pitch-shifting, and so on to tailor the audio to the user's needs,

5.      For Cognitive assistance application can have an inbuilt dictionary especially for idioms, slang, and phrases used in unusual ways.

AI advancement can not only help businesses stay compliant but win customer loyalty of a diverse user base. Usage of AI can be a step towards providing inclusive experience to demographic which is largely untapped. If companies think about catering and winning trust of a new customer base, investment in accessibility will appear very small.