OCR fields recognition - ocr

Question

can somebody help me with next question, we need use OCR technology but we don't need all text, only some fields from invoices and receips, can't find what is better solution for this

Answer 1

If you want to start mobile app development, you can use the Text Recognition Service, at the core of which is OCR technology. For example, Huawei ML Kit Text recognition API, Google Firebase ML's text recognition API, etc. You can extract the required text information(invoices and receipts if you need) ) through code.
I will list main procedures for ML Kit Text recognition integration, you can also download the demo on Github.
Preparations
1). Configure the Maven Repository Address in the Project-Level build.gradle File
buildscript {
repositories {
...
maven {url 'https://developer.huawei.com/repo/'}
}
}
dependencies {
...
classpath 'com.huawei.agconnect:agcp:1.3.1.300'
}
allprojects {
repositories {
...
maven {url 'https://developer.huawei.com/repo/'}
}
}
2). Add Configurations to the File Header
apply plugin: 'com.android.application'apply plugin: 'com.huawei.agconnect'
3). Configure SDK Dependencies in the App-Level build.gradle File
dependencies {
// Import the base SDK.
implementation 'com.huawei.hms:ml-computer-vision-ocr:2.0.1.300'
// Import the Latin character recognition model package.
implementation 'com.huawei.hms:ml-computer-vision-ocr-latin-model:2.0.1.300'
// Import the Japanese and Korean character recognition model package.
implementation 'com.huawei.hms:ml-computer-vision-ocr-jk-model:2.0.1.300'
// Import the Chinese and English character recognition model package.
implementation 'com.huawei.hms:ml-computer-vision-ocr-cn-model:2.0.1.300'
}
4). Add these Statements to the AndroidManifest.xml File so the Machine Learning Model can Automatically Update
<manifest>
...
<meta-data
android:name="com.huawei.hms.ml.DEPENDENCY"
android:value="ocr" />
...
</manifest>
5). Apply for the Camera Permission
<uses-permission android:name="android.permission.CAMERA" />
<uses-permission android:name="android.permission.INTERNET" />
<uses-feature android:name="android.hardware.camera" />
<uses-feature android:name="android.hardware.camera.autofocus" />
2. Code Development
1). Create an Analyzer
MLTextAnalyzer analyzer = new MLTextAnalyzer.Factory(context).setLanguage(type).create();
2). Set the Recognition Result Processor to Bind with the Analyzer
analyzer.setTransactor(new OcrDetectorProcessor());
3). Call the Synchronous API
Use the built-in LensEngine of the SDK to create an object, register the analyzer, and initialize camera parameters.
lensEngine = new LensEngine.Creator(context, analyzer)
.setLensType(LensEngine.BACK_LENS)
.applyDisplayDimension(width, height)
.applyFps(30.0f)
.enableAutomaticFocus(true)
.create();
4). Call the run Method to Start the Camera and Read the Camera Streams for the Recognition
try {lensEngine.run(holder);} catch (IOException e) {// Exception handling logic.Log.e("TAG", "e=" + e.getMessage());}
5). Process the Recognition Result As Required
public class OcrDetectorProcessor implements MLAnalyzer.MLTransactor<MLText.Block> {
#Override
public void transactResult(MLAnalyzer.Result<MLText.Block> results) {
SparseArray<MLText.Block> items = results.getAnalyseList();
// Process the recognition result as required. Only the detection results are processed.
// Other detection-related APIs provided by ML Kit cannot be called.
…
}
#Override
public void destroy() {
// Callback method used to release resources when the detection ends.
}
}
6). Stop the Analyzer and Release the Detection Resources When the Detection Ends
if (analyzer != null) {
try {
analyzer.stop();
} catch (IOException e) {
// Exception handling.
}
}
if (lensEngine != null) {
lensEngine.release();
}
You can also refer to this medium article here.
For more, see official documentation.

Answer 2

there are a couple of solutions that I am aware of from the LEADTOOLS SDK. Both the Master Form editor and the Document Analyzer sound like they could be useful to your use case. In the interest of full discloser, it is the company that I work for, but I would be remiss not mention it as a possible solution for your scenario.
This is a link to a YouTube video describing how the Master Forms Editor technology works.
https: //www.youtube.com/watch?v=wo6TGcdrtb4
It essentially allows you to define pre-set zones on various forms, and then compare the document to the forms in the repository. Then it will extract the data from those zones and you can manage the OCR'd data however you would like.
As well as a link to Online Documentation that shows the coded implementation of some of the functionality:
https://www.leadtools.com/help/sdk/v21/dh/to/steps-to-generate-a-master-form-and-save-it-to-a-master-repository.html
Here is a link to the other functionality that I mentioned - Document Analyzer - that one of my colleagues goes over in the Microsoft Build Post Show.
https://www.leadtools.com/blog/general/microsoft-build-post-show-v21-document-analyzer-demo/
The code for these examples, and the technology used, can be found in a 60-day free trial offer that can be downloaded from the LEAD site if you are interested in checking it out.
https://www.leadtools.com/downloads