› Modules and plugins › Document Assembly module › Classify Images plugin

Classify Images plugin

The CLASSIFY_IMAGES plugin is responsible for classifying documents using the image comparison algorithm in ImageMagick. The plugin works on the two stages of document classification:

  1. Learning: Learning is done by generating indexes for documents that are used to classify documents. See Train Transact for document classification for more information.
  2. Classification: Learnt data is used as a reference for classifying documents. When an uploaded document goes through classification, the plugin superimposes the new image over the learnt image and generates a confidence value.

Configuration

  1. Go to Batch Class Management > [batch class]  > Modules > Document Assembly.
  2. Add the CLASSIFY_IMAGES plugin to the Selected Plugins column.
  3. Click Deploy.
  4. Go to Document Assembly > CLASSIFY_IMAGES.

    The following configurable properties are available.

    Configurable property

    Value type

    Options

    Description

    Classify Image Switch

    Switch

    • ON

    • OFF

    Set this property to ON to enable the plugin.

    Classify Image Max Results

    Integer

    Any numerical value

    Set this property to the maximum number of results classified from the uploaded images that will be stored in the batch.xml file.

    Classify Image Comparison Metric

    String

    N/A

    This property is used to compare the learnt images with the uploaded images.

    Classify Image Fuzz Percentage

    Integer

    N/A

    This property is used to fuzz the distance approach in ImageMagick.

Plugin workflow

  1. You need to learn documents for the batch class prior to starting the classification process.

    See Train Transact for document classification for more information.

  2. Upload documents to the batch class.
  3. After the import process for a batch has completed, the CLASSIFY_IMAGES plugin starts in the Page Process module.
  4. The plugin classifies the uploaded documents using image-based classification.
  5. Information is written to the batch.xml file.