AUTO CROP DETECTION

Whether scanned using a flatbed scanner, or a camera based system, digitized photos usually have a margin around them. The margin color and its geometry vary depending on the scanning technique employed.

Detection of the crop region for a margin is a harder problem than it first seems. One can quickly propose simple algorithms checking for dark/light pixels for example. However such simple approaches will fail on real life samples. The shady colors of the margin, small unexpected artifacts on the sides will often produce unexpected results.

Noting the differences in the scanning techniques, BatchCrop uses three different approaches for three scenarios:

Dark Margins : This is typical for camera based scanners with projection technology. Margin has a dark color.

Light Margins : This is the case when you use a flatbed scanner to scan photos. Margin has a light color, and is usually tilted due to manual placement.

Document Margins : Similar to light margins case, but the content is a document rather than a photo. In this case, preserving page numbers for example may be an issue.

There are three buttons in the actions panel for these modes. You can also show/hide these buttons from "Options | Display".

BatchCrop will analyze the image and suggest a crop rectangle when any of these buttons are pressed. In case you want better control, there are more parameters to guide the auto crop detection :

Specified Size : If you already know the crop size, but the location of the margins vary, you may choose this option. BatchCrop will auto detect the crop and will adjust it to the specified size using the specified anchor.

Limit Search : This option limits how deep inside (from edges inwards) an auto detected crop can be. It will prevent over-cropping.

Adjust : This option will allow you to make inwards/outwards adjustments on the auto detected crop. Hence, you may discard zigzag scissors cuts for example.

Slide Mode : This option is used when searching "dark margins". Specifically developed for scanned slides, this option uses an enhanced algorithm to identify the borders. It will usually lead to smaller crop rectangles assuming the dark margins have strong contrast.

Modulo Size : This option is used when searching "document margins". It is the width of a typical "dot character" in the scan. Setting it correctly will improve document margin detection.