What are some steps we could take to reproduce the issue?
Attempted to “add annotation” to encounter. When drawing annotation box, even when setting it to the entire image, you can only get up to half of the image within the annotation, after saving the new annotation.
If this is a bulk import report, send the spreadsheet to services@wildme.org with the email subject line matching your bug report
This is the size of the box, even when viewing during the matching process, that is the most coverage I can get for the grouper in the image, so it is not just a visual bug.
Definitely a bug. We have seen this one before where the image and desired annotation are almost the same size. I have filed it as WB-1978.
Full disclosure: because we are shifting development focus to Codex, which has an entire rewrite of manual annotation, we are not likely to fix this bug in Wildbook. Rather, you’re likely to be migrated to Codex sooner than we can get to a fix.
For now, please use this workaround.
In your favorite graphics package, open the image and create blank background padding around it (see below).
Upload the new image to the Encounter.
Refresh the Encounter page a few minutes after upload, looking to see if our detector found the fish. If not, draw the manual annotation. Drawing the manual annotation in dimensions significantly smaller than the total image dimensions should allow for a more accurate bounding box.
