Will that cell have don’t cares and how will the end output show if a landmark is detected or not
Hi skylum,
If the probability of detection of an object in an anchor box is lower than the threshold, the other values of the related output vector do not matter (don’t care).
The output is a list of bounding boxes along with the recognized classes. Therefore, if there are no objects, there will be no output.
@reinoudbosch , I am referring to landmark detection where the output is if there is face in the image then if face detected then n number of landmarks (x,y) coordinates. So, let’s say a face is detected but out of 20 landmarks only 10 were visible , but the output will contain (x,y) values for all landmarks . So, how can we know which landmark is detected or not (since a landmark doesn’t have additional field of found or not) ?
Hi Skylum,
Your question seems to go beyond the scope of the current course. If you want to dive into this, you can find an interesting review here.
@reinoudbosch ,thank you ,will check it out.