getEnableAdaptiveBinarization

public final boolean getEnableAdaptiveBinarization()

Indicates whether the black and white reduction threshold applied before mark detection is derived from the image, rather than using a fixed value. Mark detection operates on black and white data, so a color or grayscale image is reduced first. The fixed threshold treats anything lighter than itself as blank, which discards marks made with light ink — for example gray shading or a soft pencil — before detection runs. The resulting field is reported as empty rather than as a low-confidence mark. When enabled, each OMR field is instead measured against the paper level found within that same field, so ink appreciably darker than its surrounding paper is preserved however light it is. Because every field is judged on its own, this also holds up on tinted or aged stock and on pages that are lit unevenly, where no single threshold would suit the whole image. Enable this for forms whose marks are not reliably solid black. If the fields cannot be measured, the fixed threshold is used, so enabling this cannot give a worse result than leaving it off. Erased marks are recognized and reported as not filled: an erased bubble keeps light residue with the printed character showing through it, which differs measurably from a deliberate mark — shading a bubble either darkens it deeply or covers its character. Marks cancelled by other means, such as crossing out or blotting, are not distinguished from deliberate marks; resolving those requires knowing which fields belong to the same question, which this method does not. This setting has no effect when the supplied image is already black and white, because no reduction is performed in that case.

Return

The value of the EnableAdaptiveBinarization property.