Face detection is a feature independent of head detection, where inside a subframe of the detected head system finds and detects the subject's face.
Face detection is exclusively done inside the given head sub-frame. Meaning, simultaneously, a green circle shows up around the head of the subject (head detection) and the sub-square marked with full and interrupted lines appears to mark the face area (face detection).
The system outputs predictions for head and face locations for every frame within the same model that analyses the entire image. It's important to point out that head detection does not imply that face detection will certainly occur as a result.