This project is for CENG3522 Applied Machine Learning class. Almost every country declare curfews due to pandemic. In Turkey, Curfew is just for people who older than 65 and younger than 20. Accordingly, we can use AI for detecting people who violate this rules. Our model can predict age and gender. It can detect the punitives in streets. For now, it can predict just one human face but it can be improved.
First clone the repository to your local.
git clone https://github.com/ozgurdogan646/ceng3522AgeGenderClassification.git
After that there are two options :
- Train your own model
- You can use directly our basic model
For now you need the change code for train your own model. This will change in future versions
Install data from this link : https://www.kaggle.com/ttungl/adience-benchmark-gender-and-age-classification
Then create your model.
>Faces folder must be in same directory with scripts.
First you need to install our model weights. They were not uploaded to git due to the file size.:sweat_smile:
Here is the link : https://drive.google.com/file/d/1lWCL83V4Y1exiHXAq4PaAEeHTAqZd4wd/view?usp=sharing
Extract this to your folder and delete models folder if it exist.
If you use your model :
>>> python createModel.py
>>> python ageGenderPredictor.py
Don't need to use createModel.py if you don't create your own model.
Eran Eidinger, Roee Enbar, and Tal Hassner, Age and Gender Estimation of Unfiltered Faces, Transactions on Information Forensics and Security (IEEE-TIFS), special issue on Facial Biometrics in the Wild, Volume 9, Issue 12, pages 2170 - 2179, Dec. 2014