A demo application for Android mobile devices that combines Optical Character Recognition (OCR), named entity extraction, and mobile Internet search. Photograph any newspaper or magazine article: AlchemySnap identifies any text in your photo, then extracts key people, companies, and other relevent concepts from the text. You can then browse Google, Wikipedia, and Twitter search results for the concepts identified in your photo.

Visual text recognition is provided by a custom OCR stack (which handles page orientation detection, shadow removal, etc.).

Natural language processing functionality is provided by my company, AlchemyAPI.

Author

• This OCR+Semantic Search app was written by Elliot Turner (ell_NOSPAM_iot@eturner.net).

Video

Features:

• real-time OCR from camera phone photographs taken via an Android mobile device.
• recognized text is fed through the AlchemyAPI natural language processing engine.
• NLP algorithms identify persons, locations, companies, and other semantic elements from the photographed image.
• mobile app provides access to Google search results, relevant Wikipedia pages, location maps, and more based on identified concepts.

back to top