Hi everybody,
I'm thinking of developping a mobile OCR app to detect words from mobile pictures. The purpose if only to detect what words are in the picture, the layout is not important. Also it would be use on very short texts.
I'm currently thinking of adapting tesseract to iphone and android.
I wonder if anyone has had any related experience? What are the limits etc.
Thanks!
Google Goggles does that... It takes a snapshot, reduces the color depth, scans through various contrast ratios, and exposes letters/words. It then performs a google search on what it found.
Hi, you can also try commertial SDK with better chances to get reasonable accuracy: http://www.abbyy.com/mobileocr
According to the web site iPhone and Android are supported.
You can try the WiseTrend OCR API ( http://www.wisetrend.com/wisetrend_ocr_cloud.shtml - sign up at http://www.webservius.com/cons/subscribe.aspx?p=wisetrend&s=wiseocr ). It is a web API for OCR based on the ABBYY engine, specifically designed with mobile apps in mind and great for processing low-quality images.