Skip to main content
An administrator can train field extraction when a project must be trained before operators start working on it, or when auto-learning proves insufficient.

​
Standard scenario

1

Create the Document Definition

Create a Document Definition. In its section properties, select Allow field location training.
2

Create the fields

Create the fields in the section, and select Can have region in each field’s properties.
3

Save and publish

Save and publish the Document Definition.
4

Create a training batch

Switch to Field Extraction Training Batches mode and create a new batch. Select the Document Definition you created, and then select Default variant from the list of variants.
5

Load and recognize images

Load document images and recognize them. Load 3 to 50 images for each kind of document.
6

Adjust the field positions

Adjust the positions of the fields.
7

Set document states

Select some documents, right-click, and click Set Document State → For training. Select the remaining documents and click Set Document State → For testing. Use about 60% of the batch for training and 40% for testing.
8

Train field extraction

  • (Standalone) Click Fields Training → Train.
  • (Distributed) Click Fields Training → Train to train on the machine where the Project Setup Station is installed.
  • (Distributed) To train on a Processing Station, right-click the batch and select Send for Training.

​
Projects with multiple document variants

For documents with variable field locations, create variants and train a classifier to distinguish them. For more information, see Variable field locations on documents of the same type.
1

Create the Document Definition

Create a Document Definition. In its section properties, select Allow field location training, and create the fields with Can have region selected.
2

Add section variants

Add section variants using one of the following methods:
  • Create variants manually: on the Data Sets tab of the section properties, click View, then Add.
  • Load variants from a database: on the Data Sets tab, click Set Up, and select Database as the data source.
  • Create variants using a script: on the Data Sets tab, click Set Up, and select Script as the data source.
3

Save and publish

Save and publish the Document Definition.
4

Train a classifier on the variants

  1. Switch to Open Classifier Training Batches mode and load document images into a new batch.
  2. Assign a reference class to each document, using variants as separate classes. Click Set Class → Add → Add, select Specify variant, and then select a variant from the list.
  3. Train the classifier with Classification Training → Train.
You can also create a classifier in FlexiLayout Studio and import it. For more information, see Export and import classifiers.
5

Train field region detection for each variant

  1. Switch to Field Extraction Training Batches mode and create a new batch.
  2. Select your Document Definition and a variant to train.
  3. Load and recognize 3 to 50 images per document kind, adjust the field positions, and set document states (For training / For testing).
  4. Train the fields:
    • (Standalone) Click Fields Training → Train.
    • (Distributed) Click Fields Training → Train to train on the machine where the Project Setup Station is installed.
    • (Distributed) To train on a Processing Station, right-click the batch and select Send for Training.
To let operators check and change the variant assigned to a document, display the variant IDs on the form with a service field. For more information, see Enable operators to change variants.
Configure auto-learning for field extraction so FlexiCapture learns to extract fields automatically as operators work on the configured project.

​
(Distributed) Send training batches to a Processing Station

(Distributed) Because training can take a long time and consume significant resources, the administrator can train batches on a Processing Station. (Distributed) Before sending a training batch to a Processing Station, make sure that:
  • (Distributed) At least one Processing Station has been added on the Processing Server.
  • (Distributed) The project has been uploaded to the server.
(Distributed) To send a training batch, right-click it and select Send for Training. After the batch is sent, a training task appears in the task queue in the Processing Server Monitor, and training runs on the Processing Station. You can send multiple training batches, and you can cancel a training task in the Processing Server Monitor. Training batches sent for training are locked until training completes, so you cannot change them during training.