Kofax Getting Started with Ascent Xtrata Pro - Manual (Page 404)

Project Builder User Interface
Ascent Xtrata Pro User's Guide 385
Forms
If this option is selected, the classifier uses the entire region of the image. This
should be used for forms and other types of documents that have a fixed layout
over the entire region of the image.
Advanced
Click “Advanced,” to display the following options:
Image Preparation
By default, ”Enable skew tolerance” is selected. When selected, the layout
classifier can internally correct for a certain amount of skew in the image. If you
are using another application that corrects skew, such as VRS, there is no need to
select this option.
Training
Use the text field to define a maximum number of training documents allowed
for a class. If the number is attained, additional documents will not be copied to
its training set. To set the homogeneity of the documents use either the slider or
enter a value to the text field.
• Max. samples per class - The Layout Classifier supports an unlimited number
of samples per class. If the sample images are very different, the Layout
Classifier internally learns different patterns for each sample. For performance
reasons, you might want to limit the number of sample documents that are
used for feature extraction. A value of 0 means no limitation.
• Class homogeneity - This feature controls how sensitive the classifier is to
variations in the layout of the images in the training set. If the sample images
are very different, the Layout Classifier automatically creates internal patterns
for each new type. These types are not visible to the user.
The greater the number of types, the better the classification accuracy but the
slower the classification speed. The value set by this control is a threshold,
which determines when new internal types are created. The higher the
number, the more sensitive to variations the classifier becomes. In most cases
the default value of 80.0 works the best.
Noise Filter
Use the slider to set the value for the noise filter.
This feature controls how to match regions with low contrast (for example,
images with a fine background pattern).
A value closer to the “max. precision” side makes the engine more sensitive to
this type of noise, and it therefore would not classify images with low contrast.
This means that even documents from the training set would not have 100%