Develop Reference

How can I be output Android only numbers in OCR?

I am developing an Android OCR.
Using tess-two in the Android made ​​the OCR.
I have downloaded the 'traineddata' file. And it has succeeded in output in English . But I want to output a number.
In the Internet gave me inform the sentence 'tesseract image.tif outputbase nobatch digits' and it has advised me to insert the generated files.
But I did not understand what I'm saying .
Please tell me the easiest way.

You will need to set tessedit_char_whitelist variable, as follows:
baseApi.SetVariable("tessedit_char_whitelist", "0123456789");
See Android OCR detecting digits only using popular tessercat fork tess-two or extracting numbers from Bitmap in android using tess-two library.

Read correct OCR data(having less accuracy) using Tesseract

Hi I have to develope an application to read OCR data .
After Googling I found that i can achieve it using Tesseract .
I got Tesseract source code from https://github.com/rmtheis/tess-two
I got its libraries by donloading NDK from http://dl.google.com/android/ndk/android-ndk-r9-windows-x86.zip
Now extracted it suppose I have extract it into E:\Software\android-ndk-r9 so I will set this path on Environment Variable
Right Click on MyComputer->Property->Advance-System-Settings->Advance->Environment Variable-> find PATH on second below Box and set like path like below picture
Now open cmd and go to on D:\Android Workspace\tess-two like below
Now downloaded other source code also from https://github.com/rmtheis/tess-two , and extract and import it and give it name OCRTest, like in my PC which is in D:\Android Workspace\OCRTest
So I have compile both two projects successfully by taking help from best OCR (Optical character recognition) example in android
But my problem is when I click on capture button then it is producing wrong text .
If anyone had done this correctly then please help me...
Thanks a lot in advance....

Do check abbyy OCR its better accuracy than Tesseract
ABBYY offers its compact OCR Engine for Android
The latest Android smart phones have
* good auto focus cameras and
* more than 3 megapixel
* so they are perfectly equipped to take sharp pictures of documents that contain textual information.
Implementing OCR on Android based applications can turn the phone in a information/text capture device. The data can be
* processed directly on the phone, for example add a new contact to the address book
* sent to a sever to get more information
* sent to a text to speech engine
* sent to a translation service
New in V4 R5: Java Native Interface Wrapper (JNIWrapper) is included!
New in V4 R7: Barcode Recognition is included!
Pricing Details http://ocrsdk.com/plans-and-pricing/
for more --> http://www.abbyy-developers.eu/en:mobile-imaging-sdk:start

How to create an MS Word document within an Android app

I want to create a word document within an android app and send that document through mail.
Is there any tutorial for creating a word document in android. I have gone through several other questions on this website but i didn't got a clear answer.
can we do this on Android using Apache POI.
If any any sample example exist then please mention.
Thanks in advance.

You can use any Java library in Android, so I do think this would be the way you could accomplish what you want (using Apache POI).
You can send the attachment by adding it as an extra to the Intent you use to create a mail message (lots of examples of that).

Apache POI looks your best bet, but note that the component that deals with Word docs only supports simple files:
HWPF and XWPF for Word Documents
HWPF is our port of the Microsoft Word 97 (-2003) file format to pure Java. It supports read, and limited write capabilities. It also provides simple text extraction support for the older Word 6 and Word 95 formats. Please see the HWPF project page for more information. This component remains in early stages of development. It can already read and write simple files.
We are also working on the XWPF for the WordprocessingML (2007+) format from the OOXML specification. This provides read and write support for simpler files, along with text extraction capabilities.
You should seriously consider whether you can use a different format for your emails - plain text, or maybe HTML.

Convert I18n Iphone Files to Android

I have an iOS App localized (with Localizable.strings files).
I've built now an Android version. I understand the way the i18n is made on android but I would like to avoid redoing the i18n, as almost all the Strings have the same meaning between iOS and Android.
Are there any tools that will do this conversion, or at least speed up the process?
Edit:
Ideally I would like to find a method/tool that would do the following
Input:
'iOS Localizable.strings files-(en,fr,it,pt,etc)'
'res/values/strings.xml'
Output:
'res/values-en/strings.xml'
'res/values-fr/strings.xml'
'res/values-it/strings.xml'
'res/values-pt/strings.xml'
'res/values-etc/strings.xml'

There is a specialised service here: Loco.
There are threads here on SO which deal with conversation:
Are there any tools to convert an Iphone localized string file to a string resources file that can be used in Android?
Any tool to convert Android's XML localization to iPhone's .strings file?
Another possiblity would be to use an online translation service. These services accepts text not only in Android and iPhone but also in many other formats. Upload the android text and download the iPhone text. A list of available services can be found on SO, see Any collaborative tool/website to localize an Android app? Some of these services are commercial, but even these provide a 30days test period. But note that the main focus is on translation and not on conversation.

Here is some code that does that: https://github.com/wrapp/LocalizedStrings2Android
Put it in a jar and it is ready to use.

You can use this tool I created on JSFiddle:
https://jsfiddle.net/danielgindi/x9njj9gj/
[StackOverflow requires me to write some code here because of the JSFiddle link]
Just paste in your localization strings inside the correct box, and you'll get the converted version in the other boxes (Between Android, iOS, and JS).
This tries to preserve comments too.
Cheers!

Android: How to match font with image document?

I'm building an OCR. I have already binerized image. But I need to know how to match font with images. I have come to know about tesseract. But it is a built in tool. Actually I need to know what is behind algorithm for matching image text with a font in .ttf format. If tesseract is the only choice for android then would you please describe some steps for integrating with windows7 as I'm not clear from Gautam's Blog. If there is any other built in method for android that match image pattern with a .ttf file please suggest me. Thanks in advance.

You'll have to train the tesseract-enginge to your font. There is an exhaustive tutorial on this topic on the projects website. You don't have to train tesseract on an android device, but you will have to deploy the training results to it.