Wish [SOLVED] Google Play Vision Services or Android Vision API wrapper

Hi,

I wish we could have a wrapper for Google Play Services related to Vision API. CMIIW, the ones that available in this forum are wrappers for samples of how to use the Vision. Thus we lost the possibilities to achieve many features. There is already a wrapper Google Play Game Services which is great. Please, create for Vision API.


Thank you,

Hi,

Today, I just created an app using Google Play Service Vision in Android Studio. The text detection is done by processing the image produced by intent of ACTION_IMAGE_CAPTURE. The result is impressive compared to just-in-time detection using the API's CameraSource. I am unable to use this gms.vision in eclipse; which I still prefer to use than Android Studio. I really wish B4A could have the wrapper for this since B4A is an awesome Android development tool that really boost Android developers productivity.


Thanks,

Why not use OpenCV? You can probably do the same things but much more!
See here:
https://www.b4x.com/android/forum/threads/opencv320-for-b4a-v1-0-released.79816/

Hi, moster67

Thank you for your response.

Yes, I am aware OpenCV and also Tesseract. From what I had learned, OpenCV and Tesseract can be complimentary in text detection: OpenCV does pre-processing to an image and hands off the result to Tesseract for text recognition. Google Play Service Vision offers simpler usage and without increasing app size cause the API is built in Google Play Services. The speed and detection result are good. Whenever pre-processing is needed I can just let the camera app in the phone to do that; so far I've never had a need to do that. Perhaps, Google Play Service Vision is just OpenCV + Tesseract. I don't know.


Thanks,

Hi again,

I found a solution by writing a wrapper library for my modified code written in Android Studio; thanks to tutorials and example code in this forum. This wrapper only performs text recognizing/scanning on an existing image file. It does not display anything.

BSTextOCRMin 1.0

Methods

• Initialize(prmEventName As String)
• Run(prmFileName As String)
  prmFileName should contains full file name, example value:
  B4X:

  fname = File.Combine(File.Combine(File.DirRootExternal, "Pictures"), "tmpscan.jpg")

Events

• onSuccess(prmBlocks As String, prmLines As String, prmWords As String)
  They are the results in comma separated. Note: They may contain empty values.
• onError(prmErrMsg As String)
• onDone

Dependencies
Permisions

• android.permission.READ_EXTERNAL_STORAGE

Installation

• Extract "BSTextOCRMin.jar" and "BSTextOCRMin.xml" to B4A Additional Libraries folder.
• Extract "add-res" onto project's folder.

Additional code
Put it below "#Region Activity Attributes".
"add-res" is a folder containing a few files required and the name might be changed to whatever you'd like. This folder should be placed in the same folder with "Files" and "Objects".

Recommended implementation

• Call camera intent to get a picture file.
• Enhance the picture (resize, sharpen, contrast, greyscale, or other) to expect better scan accuracy.
• Scan the picture with this library and get result.
• Optionally, delete the picture file after done.

I share this for B4A community to be used freely. Comments are welcomed. Thank you.


Cheers,

You should post libraries in the libraries forum.

Hi, Erel

I copied to https://www.b4x.com/android/forum/threads/minimal-text-ocr-using-google-vision.83761/


Thanks,