During the prediction the trained IKOSA AI App will assign an ID to each and every found instance. These IDs not only determine the order in which the instance are listed in the csv or xslx output, they can also help finding the instance within the visualization. This page documents the peculiarities of instance numbering within the visualizations.
Order of the instances on small images
The instances on small images (smaller than a “tile“, more on this under the next point) are ordered by a line-dominant scheme. This means the image is read line by line and from left to right when numbering the instances. The relevant point for the instance is the upper left corner of its bounding box. Simplified, this can be imagined like this:
Order of the instances on large images
Large images are split into “tiles“ within the application. On each of those tiles the numbering is line-dominant as described above. However, neighbouring tiles are numbered in sequence from left to right and top to bottom. This can result in a pattern like the one below.
The default tile size for IKOSA AI Applications is 2048 x 2048 Pixels. So, if your image is larger than 2048 Pixels in any direction, expect your instance numbering to follow this presented pattern.
Numbers are omitted on small instances
When labelling small instances, at some point the instance id overlaps with too much of the instance - basically hiding the instance behind its own index visualization. To avoid this, a visualization threshold for the instance number exists. By default, if the numbers' bounding box is more than 5 times larger than the instance area, the number is not displayed.
Limit on number of instances to be numbered at all
When predicting instances on very large images or WSI, the predicted number of instances can be more than can be reasonably analyzed visually. And as the instances with very large indices would need to be very large to actually have a displayed number (see previous point) there is a maximum number of instances until when the instance numbers for all instances are displayed.
By default, this number threshold is 100.000 instances. So, if your prediction contains more than 100k instances, no instance number will be shown in the visualization at all.