B4A Library [B4X] TextRecognition based on MLKit

This is a B4A + B4i solution. It recognizes text in images based on Google ML Kit.

Dependencies:

- B4i (local Mac only): https://www.b4x.com/b4i/files/MLKit.zip
The files and folders should be unzipped into Libs/Extra/MLKit

Don't miss:

- B4i dependencies + #PlistExtra (for MediaChooser) in Main module.
- B4i bundle files under Files\Special. The example includes bundles for Latin and Chinese. Other bundles are available here: https://www.b4x.com/b4i/files/MLKit_Bundles.zip
- B4A dependencies in Main module.
- B4A manifest editor. 2 snippets for MLKit and one snippet for FileProvider (MediaChooser).

Usage itself is very simple. See the example:

https://www.b4x.com/b4i/files/TextRecognition.zip

Updates:

-v1.11: Dependencies in B4i updated to v10+ build-chain. Note that the local Mac dependencies were updated.
- v1.10:
Support for non-Latin languages added. There are 5 language models: Latin, Chinese, Devanagari, Japanese, Korean
Recognizing non-Latin languages requires some configuration:

• Initialize the recognizer with the correct model.
• B4A: add reference with #AdditionalJar: com.google.android.gms:play-services-mlkit-text-recognition-chinese (replace chinese with the relevant lowercased model name).
• B4i: add the bundle of the relevant model: https://www.b4x.com/b4i/files/MLKit_Bundles.zip

Note that B4i MLKit package for local mac was updated as well as the #AdditionalLib list.

 

[personalapps]

Was amazed to see it recognize even text which was in dark !

 

[scsjc]

impressive, it works almost 99% fine, when placed at a very bad angle and vice versa ???

 

[benyamin]

I used these codes in the B4A project, but it gives this error. It seems that it must be a cross-platform project. Thank you, can you fix this problem?

@Erel

 

[Erel]

The ML Kit libraries work with B4XPages. Don't waste your time with activities.

 

[Erel]

V1.10 released with support for non-Latin languages. See first post for instructions.

 

[byz]

i do,but got error.this is my test:
B4A:12.8
my pc:windows11 21H2 22000.2538
Android emulator：android 7.1.2
my phone：harmonyOS 4.0.0

#AdditionalJar: com.google.mlkit:text-recognition
#AdditionalJar: kotlin-stdlib-1.6.10
#AdditionalJar: androidx.arch.core:core-runtime
'#AdditionalJar: com.google.mlkit:text-recognition-chinese  'error
'#AdditionalJar: com.google.mlkit:text-recognition-japanese  'error
'#AdditionalJar: com.google.mlkit:text-recognition-latin  'error
'#AdditionalJar: com.google.mlkit:text-recognition-devanagari  'error

'ok but ocr nothing  
#AdditionalJar: com.google.mlkit:text-recognition-korean

korean log in b4a ide: com.google.mlkit.common.MlKitException: Failed to run text recognizer text-recognition

 

[personalapps]

I am also getting error for devanagari

Not perfect but works:

In main: #AdditionalJar: com.google.android.gmslay-services-mlkit-text-recognition-devanagari

In B4XPage_Created: recognizer.Initialize("Devanagari")

 

[Erel]

my phone：harmonyOS 4.0.0

It will not work on devices without Google SDK.

For further discussion please start a new thread.

 

[Erel]

Dependencies in B4i updated to v10+ built-chain. Note that the local Mac dependencies were updated.

 

[Erel]

B4i Mac dependencies were updated with a fix to this issue: https://www.b4x.com/android/forum/t...uilding-on-hostedbuilder.170002/#post-1040963