Search code examples
pythonazureface-api

Minimum faces required in a Person Group - Azure Face Api

I just came over Microsoft Azure Face-API cloud-based service for enabling face recognition in my python based application. But according to my previous experience in developing Face Recognition apps, my models used to require at least 3-4 persons to classify faces correctly(to some extent).

My question is that is there any such minimum required persons that are needed to be added in a personGroup so that model can be then trained to classify faces correctly.

I just wanted to know this before I make a hasty decision of opting the Azure Face API as my primary FR platform.

Solution

  • As I known for Azure Cognitive Services about Face detection and recognition, there is no minimun requirement for persons in a image. However, there are three similar services below for different scenarios that I think you must to know. The choice to use one of them is up to you.

    1. Face detection with Computer Vision. I think it's the simplest solution with out of the box for using easily on Azure.
    2. Azure Face API for Face detection and Face recognition, also without pre-trainning like the first one.
    3. If these above two can not satisfy your needs, Azure Custom Vision Service can support you to train your own model.

    They all have the Python samples in their QuickStarts topic that you can refer to. Now, I think the only question is which one is you real needed.