Deep CNN Face Recognition App

This project trains deep learning models on databases with human faces to be recognized. These models are afterwards converted in a lite format and implemented on Android to measure accuracy and time

Project Structure

Android

• Contains the Android application used for facial recognition
• Created with Android Studio
• Ability to choose deep learning model and face database to test from a list
• Select photo from gallery or take one with the camera
• Measure accuracy and time

Python

• Contains .ipynb script used on Google Colab to train models
• Contains archive with the face databases used

Deep learning models used

• MobileNet
• EffNet
• L-CNN

Face databases used

• Essex
• Jaffe
• ORL