Diction

** Made for Hophacks Fall 2018, but they are using HackerEarth **

Inspiration

70-80% of people with poor reading skills are likely dyslexic. One in five students, or 15-20% of the population, has a language based learning disability. Dyslexia is the most common of the language based learning disabilities. Diction is a tool meant to assist those struggling with reading.

What it does

Using Google's Optical Character Recognition (OCR) and Princeton's Wordnet, Diction can read text from your camera feed as well as define words and find complex synonym relationships.

How I built it

Princeton's Wordnet data to generate over 100k word definitions and complex relationships between them. Python scripts to parse Wordnet datasets. Google's Optical Character Recognition (OCR) API to detect text from camera feed with ML and CV. Android Studio + Java to bring it all together.

Who does this impact?

-Diction impacts a wide selection of people. It helps people with dyslexia distinguish and understand words and sentences that would otherwise be a challenge. Diction features a word tracking feature that replaces natural text with text that is formatted and easier to read. Children can also benefit from this by using Diction which removes the daunting aspect of reading as well as expands vocabulary with definitions and helpful synonyms.

How to use Diction?

Point camera towards reasonably legible text. Single tap a word for text to speech word, definition, and suggested synonym. Double tap a paragraph for text to speech. How do I get Diction?

For now, it's only available on android devices! Visit the demo github link and download the .apk file on your android device. With the right settings on (it will tell you), it will prompt you to install after download. https://github.com/justinqle/Diction/blob/master/release/v1.0.apk Resources:

Google's Vision API - OCR: https://cloud.google.com/vision/docs/ocr Princeton's WordNet - Dictionary/Thesaurus: https://wordnet.princeton.edu/