I have two IP cameras placed at gates. CheckIN and CheckOut Cameras video feed will be fetched though RTSP. Whenever an unknown or unregistered person comes, it should mark them UNKNOWN and if I click on Register, then his personal data should be filled and 500 images to be taken and system to be trained. The personal data of individuals should have EmpID, Name, Department, ValidUpto fields. This should be stored in the database. Also tables should be there for the days attendance which should have, CheckIN and CheckOut timings of the individuals.
The face recognition system should not given false positive results. It should be robust and accurate.
Preferred envt and modules :
Face Detection : openCV DNN/ Deep learning/ Any other than HaarCascade since it gives very much false detection.
Face Recognition : Deep Learning which should define the layers and create model after training. more images can be taken to increase the accuracy. Should not detect any unknowns with the registered employees (No false positives).
Backend- Sqlite/ MySQL/ Any
Interface - preferably Web based interface. interface preferably should be as per attached file or better.
The system will work on the standalone PC with NO internet connectivity.
***** Face recognition expert ***** Hello. I have done lots of face recognition project in the past. I can show you my past project demo. Please contact me and discuss more detail. Best regard. Sandeep.