Deepface : Building And Training A Custom CNN For Face Recognition

Abstract

Face recognition systems are utilized in various applications, ranging from security to personal devices. However, modern systems require additional efforts after identifying individuals. This research presents an advanced face recognition system with five main objectives: Liveness Detection, which ensures the model can differentiate between live subjects and still images or videos; recognition of Various Faces, equipped with live detection for simultaneous identification in crowded areas; a robust model that maintains high accuracy despite challenges like varying brightness and obstructions; Expression analysis for real-time feedback, providing sufficient display services through an API to integrate seamlessly with various software platforms. The system employs deep learning techniques, focusing on convolutional neural networks (CNNs) and computer vision.