• The Optical Character Recognition (OCR) model identifies and extracts information from digital images and PDFs. Refer to the Understand OCR model page to learn more.
• You can build OCR models suited to your business needs, or choose a ready to use model that can be directly deployed in your applications. 

To set up an OCR model, follow these 4 steps:

Step 1: Create an OCR model

Step 2: Add training data

Step 3: Verify the model summary, train and test model

Step 4: Publish and use model

Step 1: Create an OCR model

1. Navigate to Microservices under the DEVELOP section in your Creator homepage. All your microservices will be listed here.
   

2. To create an OCR model:

• If you're creating an AI model for the first time, click the + Create New button in the center of the Microservices page.
• If you've already created AI models, then click the + Create New button in the top-right corner of the Microservices page.

3. Click the Create button under AI Models. The AI Models home page contains two model types: Custom models that can be built according to the needs of the user and Ready-To-Use models that can immediately be deployed into their applications.
   Note : You can create both custom and ready-to-use OCR models. To learn more about the model types, refer here.
   

4. Click the OCR card under the Build-Custom Models section in the AI Modeler page.
   
5. Enter a Model Name and select the type of Training Data (Bills, Cheque, Forms, Invoices, etc) that you're going to use to train the model from the given dropdown menu. If you choose the Other option, you need to specify what type of image you're going to upload.
   Note : The model name cannot exceed 30 characters in length.
    
6. Click Create Model. The Add Field screen will open.

Step 2: Add training data

Training data is the primary dataset that is used to train a model, so that it can perceive the input information correctly and make accurate decisions based on the information provided. This ensures that the model performs the way it's intended to. For the OCR model, a set of images or PDFs of the same layout needs to be added and the values to be extracted from them must be tagged, so that model will know which values are required to be extracted from the training data added. You can check the guidelines section for more details.

• Add fields and data types 
• Add images
• Tag fields

Add fields and data types

1. In the Add Field screen, enter the Field Name of the field whose value you want to extract and choose its relevant Data Type : Text, Number, or Date. 

2. Click + Add Field to define more extraction fields and their field types similarly. Click Next at the bottom when you've finished adding the required data.
   
   

Add images or PDFs

Tag Fields

After adding the fields and uploading the required images or PDFs, you need to tag the added fields to their corresponding values in each of the training data that you've uploaded. This helps the model to learn and identify which values must be extracted from the training data. All the data that can be extracted from your training data will appear highlighted. You can now choose to extract the specific text portions for your business needs.

1. Click on the field value in an image or PDF page and tag it to its corresponding field. You need to click and drag completely over the values that need to be tagged.
   
   

2. Click Tag Next and repeat the repeat the previous step in all the other uploaded images.
   

3. Click Next after the mapping is over. The Model Summary page will open.

   Note : Unless you've tagged all the added fields in all the images, you cannot proceed further.
   

Step 3: Verify the model summary, train and test model

After adding the training data, you can review the model details, such as Model Name, Model Type, type of Training Data, and the Number of training data (images or PDFs) added. If you need to make any modifications, you can go back and make them. Otherwise, you can proceed to train the model. 

• Train model
• View and manage model details
• Test model

Train Model

Before you can actually use your OCR model in your application, you have to train it to perform the way you want.

1. Check the details of your model in the Model Summary page and make any necessary changes by going Back. You can modify the model name, type of training data, upload additional training data or remove unwanted data.

2. Once you've made the necessary changes, click Train Model.
   

   Note : Model training may take some time, so you can either stay on the page and wait, or you can close the page and come back later.

View and manage model details

After the training is complete, the user can view the status of the model (trained, failed, and draft), the model type, the date it was created on and updated on, and other details as mentioned below. 

• Model Details
• Version Details
• Model Deployment

Model Details

Under this section, you can view the current version of your model, the type of training data, and the field names whose values need to be extracted.

Version Details

In this section, you can view the number of versions the model has, what version the model is currently running on, the model creation date, the number of fields added, and the count of trained data.

Note : To learn how to create a new version of your model, click here.

Model Deployment

In this section, you can view the App Name, Form Name, and the Field Names in which the model is deployed in. You can also filter between different environments to check which environment a model is deployed in.

Test Model

After training, you can test the model's reliability before deploying it in any of your applications. This ensures that the model identifies and extracts the required values correctly.

1. Click Test Model in the top-right corner of the page that appears after the model is successfully trained. 
   

2. Upload an image or PDF in the Test Model screen that appears. The model will try to display the extracted values on the right under Model Output.
   
1. Note :
   
1. The important criteria while extracting values is the position of the added fields along with their tagged values. Click here to learn more.
   
2. If you've trained your model with images of slightly varying layouts and upload a test image that falls under one of those layouts, then the values will be extracted correctly.
   
3. If you've trained your model with images of slightly varying layouts and upload a test image that doesn't fall under one of those layouts i.e it is a different image or one in which the field whose value is to extracted is located at a different position, then the values will be not be extracted correctly. The same applies to PDFs as well.
4. To improve your model output, you can retrain your model or edit the model details and train again.
   
   

Manage OCR Model

After you train your model, you need to publish it to make it available for deployment in your applications. Your users can then use your model and start extracting values from their images.

Note : You cannot un-publish the model once it has been published. You can still make changes to the model and train it again.

• Retrain model
• Publish model
• Use model

Retrain model 

Retraining the model with the additional images or PDFs of similar and slightly varying layouts helps your model identify and extract the required values more accurately. This reworking on the model's efficiency allows the model to be tuned specifically to your business perspectives.

1. Click the three-dot ellipsis at the top-right corner of your page.
   
   
2. Click Retrain to train your model again. A new version will be created and listed under Version Details.
3. Click Edit to make changes to your model. These changes include modifying your model name, modifying the fields and their data types, adding/removing training data, and modifying the tagged values. You need to train your model again for these changes to be reflected.
4. Click Rename to edit your model's name. A popup will appear, where you can edit the model's name and click Rename.
5. Click Delete to delete your model.

Step 4: Publish and use model

After you train and test your model, you can publish it to make it available to your users and start extracting the required values. After publishing, your model can be used by users in your applications. In case you don’t want your users to use the model, you can delete the model.

1. Click Publish Model in the top-right corner.
   
2. Click Publish in the Publish popup that appears on the screen.
   
   

Use Model

1. After the model has been published, you can click Use Model either in the popup that appears or at the top-right corner of your page.
   
2. Select the Application Name and Form Name from the dropdown list in the Use Model popup that appears, then click Use Model.
   
   The user will be redirected to the form builder of the application they've selected to have the model deployed in. The OCR popup will appear with the Model Input screen open.
   Note : The trained and chosen model will already be selected in the Select Model section. You can choose a model when you create a ready-to-use OCR model. 

3. In the Model Input section, select the source field from the drop down menu and click NEXT.

4. Confirm if the Extracted Field and Field Type are selected correctly. By default, all the defined fields will be selected and created. You can deselect them if required.
   

5. Click ADD FIELD. A new OCR field will be created.

You can now access your app in live and upload the required images or PDF in the source field. The OCR field will try to identify the uploaded input and the extracted values will be displayed in the extracted field.

Related Topics

1. Understand AI modeler
   
2. Understand OCR model