People-tracking-with-Age-and-Gender-detection

A combination between people tracking and age and gender detection

alt text

Abstract

Combining people tracking with age and gender detection is a good idea for many and many applications in real life scenarios such as caffe store management to gather the information of customers for further analysis, or in/out people control for security purposes in buildings ...
This is just a small step of putting the state-of-the-art image processing techniques together.

Firstly, faces are detected in the frame using the famous caffe model _res10_300x300_ssd_iter140000.caffemodel.
Secondly, age and gender of every person is predicted also using caffe models _agenet and _gendernet.
Each person is then assigned an ID and tracked over time, even when they are out of the frame for not so long (5 seconds), whenever they come back in the frame, their ID will remain the same.
A picture of the person is then saved with the information of him/her.

pip install requirements.txt

python detect_and_tracker.py

I can get 16fps in my Core i5 desktop with the solution of 640x480.
If you want to train your own models for age and gender detection, have a look at https://github.com/dpressel/rude-carnie.

A great thank to those who have done fantastic work