FPENet inference confidence seems low

[tom_s]

I’ve copied this question to the TAO toolkit forum… maybe someone there can help.

FPENet inference confidence seems low

note: I tried to move it to this forum from here but couldn’t , so this is a copy… (the deepstream forum didn’t know how to answer this) When running the FPENet deepstream example application on a camera input, the confidence on each landmark seems to be in-between 0.2 and 0.35 When the facial landmarks seem to be very accurate it goes to the 0.3 side, when turning profile and the hidden points are all over the place, it drops to 0.2. Is this expected or is there something wrong ? I was expec…