Seems to be due to the detection area being not fully accurate. Green vs red shows the difference between actual and detected
replyThere is an extra green square where no egg is present, so it's a false positive in the expected.
replyHi! I’m the author of this blog and benchmark. You’re right. I’ll fix it in the ground-truth dataset. Thanks for pointing it out.
replyGreat, happy that it was helpful
reply