SamiUddin Hi, Whenever I try Depthai Gen2 Facial Recognition demo on more than 3 faces the output gets stuck. Have a look at. How do I fix this. After that image is stuck here.
jakaskerl Hi @SamiUddin Seems like there is a problem with the first two manips and face detection node branch. Every face that is added lowers the fps of that branch by 10 as shown on pipeline graph. The rest of the pipeline seems fine despite adding more faces. But everything seems to crash when the problematic branch is unable to process images fast enough, which saturates all queues and blocks the pipeline from the start. I'd suggest adding some logic that clears out the queues in case of such blockage; or maybe limit the model to only 3 faces at once. Hope this helps, Jaka
erik SamiUddin We don't officially support it. If you want, you could update it, and implement something similar we have in our SDK (see script here), so anything above eg. 3 faces gets discarded. Thanks, Erik