There are several face detectors in model zoo (face-detector-retail004 & 005, face-detector-010*). They seem to be very good and fast. Is there a sample or a simple guide how to train and convert custom model so it can benefit from using the DetectorOutput layer?
There are a few of custom models that you could use as a workaround such as YOLOV3, mxnet & etc.
These custom model required to be converted into IR before using them with openvino samples/applications.
You can refer here:
Intel will no longer monitor this thread since we have provided a solution. If you need any additional information from Intel, please submit a new question.