Modifying the model to meet our requirements

[Dhruv88]

The camera at the acad block entrance is at a height. Thus we don't have a frontal view of the face. We will need to check if we can still somehow transform the face and then recognize it. To do this first we need data from that camera.

The task is to get some data from that camera so we can analyze it. For this we might need to ask ITIS.

The data can be a video clip of some duration where we have people entering the admin block.
If ITIS is asking for permission to get the data then just have a look at the footage to get an idea

[Dhruv88]

UPDATE:

1. We need a model that can detect and recognize multiple faces. Also, keep in mind the quality of face image can be low so that will need to be handled. Try to find a model that meets at least some of these requirements else we will need to develop something on our own
2. Till we get data from ITIS try to find a similar CCTV footage dataset if possible