I tested the Google ML kit Object recognition,Face detection and Image Labeling using ML Kit for Firebase (in Android app).
Followed the official documentation, and all worked well.
What is interesting about the ML kit, at least from my experience, that it avoids classifying/detecting people. Does anyone has the same experience? Is the ML kit deliberately not detecting/classifying people in images?
I tested object detection and face detection as well as labeling (including paid more precise Cloud Vision API) , and in all case the result-set contained everything but no People/Person category.
Even if the face detector correctly detected the faces, the object detector usually did not find any object, or when an object recognized, it was some non human object...
The same with the annotation labels. These labels usually contain category descriptions like wear, jacket, hat, glasses, smile... but never a person.
My intent was to add an additional intelligence to my surveillance app in the motion detection to confirm people in the image. But it looks the ML kit is not suitable here...
Any idea?
Below is the result of: -Face detect -Object detection and classification -Labeling
Faces detected, but no person recognized or labeled.
