Android application able to automatically translate text pictures captured by smartphone camera. The result image of process is similar to original, but contain the translation result in the same location of original text.
-
Text Detection: Is based in work of Menglong Zhu on Literate PR2 architecture. An Android application, with contribution of Andrei Nagornyi, is available here.
-
Text Extraction: The character recognition is executed by Tesseract OCR, created by Ray Smith. To embed it in the application was used tess-two, developed by Robert Theis.
-
Text Translation: To translate text is used MyMemory API, property of Translated S.r.l, available here (free and anonymous usage is limited to 100 requests per day. Verify the Service Terms and Conditions of Use).
Another resources documented in source code.
-
Install Eclipse IDE and ADT;
-
Install Android NDK;
-
Download OpenCV Android SDK and configure;
-
Check paths in Android.mk (jni > Android.mk);
-
Clone tess-two and check the project dependency (Project Properties > Android > Library);
A Nexus 4 device has been utilized in the tests to verify system functionalities. This article describes how to configure hardware devices to run Android applications.
The current minimum SDK required is 14 (Android 4.0 Ice Cream Sandwich).
