Some researchers prefer to include loops which are detected partially, others prefer to exclude loops which are only partially in the captured image. The ACAS default setting is to exclude partial loops. This can be overridden by adding this text in the experiment-set description:
In this example, with manual analysis, we found 9 closed loops that are fully enclosed in the captured image.
In automated analysis, seven closed loops are detected that are completely enclosed in the image, less the margin. A key distinction between the auto and manual analysis is the image margin. The algorithms are not able to accurately detect structures to the very edge of the image. The greater the resolution (increased number of pixels), the smaller the margin will be (that is, the algorithms will do a better job up to the edge. In this low-resolution image (0.65 pix/µm), the margin is relatively large.