AI-Powered Facial Analysis: Real or Fake Image Detection, Age Group Classification, Grooming Analysis and Emotion Recognition

The development of deepfake technology has made it extremely difficult to tell the difference between real and altered facial photos, which presents significant privacy and security risks. This project offers a multifaceted deep learning-driven solution that can classify age, identify emotions, and conduct grooming analysis in addition to detecting deepfake photos. The method achieves 93.91% success rate in deepfake identification using a model built on CNN that was trained on datasets of real and fake faces. Furthermore, an OpenCV-Caffe model categorizes people into various age groups, a Vision Transformer (ViT)-based beard recognition model achieves 99.9% accuracy, and an Xception-based emotion detection method reaches 98.65% accuracy. Users can input photographs, evaluate several faces, and generate a detailed report on age distribution, feelings beard presence and real/fake categorization using the system's real-time Streamlit user interface. This AI-powered solution addresses the growing worries about digital verification of identity and distorted media by offering a safe, effective and automated method of facial analysis.