DetectBlankPage and DetectBlankRectangle are similar functions that examine a given area for significant amounts of data. What is "significant" is defined in terms of object size. For example, a group of just a few pixels is probably a speck, while a group of pixels over 20 pixels in width or height is probably a deliberate mark. There is a GapFill option which runs a dilate black operation on the image during detect, which can be used to grow individual characters into word or sentence sized objects. This is useful for ignoring large noise, such as stray pencil marks, while accurately detecting data such as words.

The continuous tone detection algorithm may, in this operation, provide different results from the bitonal algorithm, even with visually identical input data. This is due to the use of a high pass filter in the continuous tone version to handle images with varying brightness and contrast. If any objects that are found exceed the minimum size, the image or region will be declared "not blank". The confidence value returned by the operation depends on the number of objects found; large numbers of objects near or over the size threshold will return a high confidence "not blank" result, while large numbers of objects just under the size threshold will return a low confidence "blank" result.

DetectBlankPage is intended for use on entire pages, and therefore its options are relative to the edges of the image. Values can be set for the margins, telling BlankPageDetect to skip over certain areas (such as hole punches that are expected in the left or right margins). BlankPageDetect also has a fast but rudimentary border detection algorithm that will skip over dark borders, so that they are not interpreted as data. Margins are relative to any detected border. If the simple algorithm used by the IgnoreBorder option is not sufficient to detect the borders of your images, then you should process the images first with the RemoveBorder method, then run DetectBlankPage on the resulting output.

DetectBlankRectangle is intended for use on portions of an image, and therefore its options are relative to a given set of rectangle coordinates. This means your images must be aligned before DetectBlankRectangle can be reliably run; refer to the topic I Need to Get Data from a Specific Area on a Form for more information. The IgnoreBorder option is available, and will run inside the given rectangle.