Facify consists of Machine Learning and Neural Network models for faces and identifying their age and various other details.
- Clone the repository using
git clone git@github.com:Facify/Models.git. - Install the required libraries and packages.
- Download the datasets require in dataset folder.
- Ready to run the code in jupyter notebook.
-
To download utk dataset go click on the following link UTK_dataset.
-
Add the unzip the dataset and copy it to the dataset folder in the repository.
-
Here are the models for UTK dataset as described below
- CNN_UTK_Model_1 is a shallow CNN model that uses regression to predict age of the facial image.
- CNN_UTK_Model_1_AgeRange is a shallow CNN model that uses regression to predict age range(range of 5) of facial image.
- CNN_UTK_Model_1_AgeRange_Classification is a shallow CNN model that uses classification to predict age range(range of 5) of facial image.
- VGG16 Model
- VGG19 Model
- MobileNet Model
- CNN_UTK_Model_1 is a shallow CNN model that uses regression to predict age of the facial image.
-
The Regression models have been observed to provide the best results for people in upper spectrum of age.
-
While the classification models provided better results for lower to medium spectrum of age.
