The
Process tab on the Ribbon is used to test the processing of documents, to perform recognition, and
to train the project.
Additionally you can configure benchmarks for separation, classification, and extraction.
Settings available in the
Train group
setting
|
Name and Description
|
|
The text for this icon depends on how your project is configured.
-
Classification /
Separation & Classification
-
Trains both classification and separation. This setting is available only when
Standard Document Separation or
Trainable Document Separation is enabled in the
Project Settings. Otherwise, this trains the classifiers, but only if a layout or
content classifier is configured for your project and the training sets contain documents for your classes.
|
|
-
Extraction
-
Trains the trainable locators, if training documents are added to the
Extraction Set.
|
|
-
Optimize Content Classifier
-
Takes the content of the selected document set and optimizes the parameters for content classification using
the Adaptive Feature Classifier (AFC).
|
Settings available in the
Test group
setting
|
Name and Description
|
|
-
Recognize
-
Performs text recognition using the default profile for the recognition settings.
|
|
-
Separate
-
Performs document separation and shows the results in the
Document Separation Results window.
|
|
-
Classify
-
Performs classification.
|
|
-
Extract
-
Extracts data using the field and locator definition of the selected class. Refine your extraction by
selecting one of the following submenu settings:
-
Extract Selected Documents
-
Extracts the data from the selected documents using the field and locator definition of the selected
class.
-
Extract Selected Documents without
Validation
-
Extracts the data from the selected documents using the field and locator definition of the selected
class without
Validation.
|
|
-
Process
-
Performs classification and extraction. Refine your processing by selecting one of the following submenu
settings:
-
Process Selected Documents
-
Performs classification and extraction on the selected documents.
-
Process Selected Documents without
Validation
-
Performs classification and extraction on the selected documents without
Validation.
|
|
- Clear Data
-
Clears the data for the selected documents. You can clear data recursively by selecting one
of the following submenu settings:
-
Clear Selected Documents
-
Clears the extraction data and classification results for the selected documents.
-
Clear Documents Recursively
-
Clears the extraction data and classification results of all documents.
-
Clear All
-
Clears the extraction data and classification results for the selected documents.
|
|
- Load Data
-
Loads any saved extraction data or classification results for the selected documents.
|
Settings available in the
Benchmark group
setting
|
Name and Description
|
|
-
Separation
-
Loads all documents in the selected documents subset into memory, merges all of the pages into a single
document, and separates this document using your existing separation settings. A progress bar indicates what percentage of the
documents are processed. When the progress bar reaches 100%, the
Separation Benchmark window is displayed and displays the
separation benchmark results for analysis.
|
|
-
Classification
-
Displays the
Classification Benchmark window and calculates a detailed analysis of the
classification quality for the
Classification Training training documents set.
-
Classification Benchmark (All Classification Settings)
-
Displays the
Classification Benchmark window and calculates a detailed analysis of the
classification quality for a
benchmark
set using
all of the project classification settings.
-
Classification Benchmark (Specific Classifier)
-
Displays the
Classification Benchmark window and calculates a detailed analysis of the
classification quality for a benchmark set using
a specific classifier only.
|
|
-
Extraction
-
Displays the
Extraction Benchmark window and performs an analysis of the extraction quality. To
refine your extraction benchmark results, select one of the following submenu settings:
-
Extraction Benchmark (Selected Class)
-
Displays the
Extraction Benchmark window and performs an analysis of the extraction quality for the
currently selected class. All documents in the selected documents subset are extracted using the extraction settings for the
selected class. The documents are not classified before being extracted. You can save the results of the benchmark in a file.
-
Extraction Benchmark (Selected Class And Children)
-
Displays the
Extraction Benchmark window. All documents in the selected documents subset are
classified and extracted using the extraction settings for the selected class and its child classes. The extraction benchmark
results however, only consider the documents that are classified as the selected class or one of its child classes.
Use this setting when there is a common base class for several child classes that contain similar
fields. For example, you have a project that processes invoices, orders, and other documents. The invoice class has several
supplier subclasses. Since the child classes have similar fields, you can perform an extraction benchmark to test the extraction
results for these classes.
-
Extraction Benchmark (All Classes)
-
Displays the
Extraction Benchmark window. Performs an extraction benchmark for all of the
classes in your project. All documents in the selected documents subset are classified, and then extracted using your current
settings.
Use this setting when you have a project that contains several classes with similar fields. For
example, an project that processes invoices has a different class for each supplier.
|