LibroBridge

Inspiration

A lot of non profits in India have shelters for the differently abled people. Because of vision loss, they are not a lot of options for them as a form of entertainment or leisure. Books can be used to fill such gaps because of their easy cost of printing and producing braille books for people who have vision loss. On the other hand audio books can also be used as an option. While this may be sound economically viable, there are not a lot of selection of boks for people. And it requires speical expertise in translating books to braille or audio books.

What it does

This is where LibroBridge comes in, closing the gap in such areas. The aim is to use existing books in the library and easily convert it to audio books or braille for printing. Inorder to achieve this the project relies on cheaply availlable android smart devices with camera. The app is basically a scanner which scans each page, does OCR and extracts the text, and converts the text to speech and braille unicode characters for printing. The final output is an audio book for the pages scanned, braille text and a normal txt file with the scanned pages.

How we built it

The app does not rely on any cloud sservices for both OCR and text to speech conversion. The app uses google ML Kit text recognition to perform OCR and extract the text and the standard text to speech engine built in with all android devices. By relying on offline methods, it serves as a demonstration on how it can be used in places with poor connectivity.

What's next for LibroBridge

Overall the app serves as a simple PoC on how to leveragee such simple technologies to solve bigger problems. In the future we can add an even more powerful OCr (AWS Textract) and TTS libraries (Amazon poly, google cloud TTS) services to better generate the text and the audio along with giving the user more control on pre processing the images (ex: cropping exact text areas) to improve the performance and the outcome.

Built With

• android
• kotlin