Face Recognition Android

An entertaining mobile application using a neural network for facial recognition.

Facial recognition works by measuring the unique characteristics of a user’s face and then comparing the captured source data with an existing database of entries. First, the app detects the picture frame with the user’s face. Then the neural network detemines user’s facial landmarks, which are a 128-dimensional array. Then the application determines the similarity between the input face landmarks and the landmarks of other users, stored in the database.

The Android version of the Face Recognition app allows for recognizing users locally, without an Internet connection. We use CPU power to quickly execute all the operations, but we can opt for using the multiple GPU cores to improve the performance.

ABCloudz utilized Java to build the Android application. We leveraged an open-source dlib library and two already trained neural networks to find the faces on the pictures and extract facial landmarks.

Be sure to check the related blog post and the video above to learn more about using neural networks to create face recognition applications.