• category Machine learning 📡
• author Juan Diego Alfonso [email protected]
• copyright GNU General Public Licence
• source GitHub

This model was created to explain how to recognice the gender of a person based in the face using CNN(Convolutional Neural network) with Keras.

You can create a virtual environment using following commands.

In linux maybe you need the following package sudo apt-get install python3-venv

This project is a CNN based in a large-scale image recognition training model, as called VGG, this model is based in layers and a compression of an image, the schema is explained in the following graph:

To train the model just run with python 3.6 and the environment created the file training.py, ensure the route of the 'datadir' variable is set as your dataset photos is. To download the dataset you can see follow URL. https://storage.googleapis.com/pretrained-model-gender/gender_dataset_face.rar

The VGG perfonrmance will be ploting and the image is one like this.

To run the recognition ensure the model is in the folder 'model' and run detection.py and comment the variable 'model_path', if you haven't the model the app download a model automatically.