Rancang Bangun Sistem Deteksi Wajah Berbasis Haar Cascade dan OpenCV: Analisis Kinerja pada Kondisi Pencahayaan dan Pose Bervariasi

Abstract

This research aims to design, implement, and evaluate an efficient two-stage computer vision system for face detection and age estimation. The first stage, face detection, uses the Haar Cascade Classifier method implemented through the OpenCV library. The second stage, age estimation, utilizes classic feature extraction (Local Binary Pattern/Histogram of Oriented Gradient) followed by a Support Vector Regression (SVR) model. This two-stage system is specifically analyzed under varying operational conditions, including changes in face pose, occlusion, and lighting intensity, to measure the impact of face detection failures on age prediction accuracy.. This research quantifies the inherent trade-off between the computational efficiency of Haar Cascade and the reliability of age estimation accuracy in non-ideal environments, providing essential data for the development of more adaptive facial biometric systems.