<![CDATA[Security system technology with identity recognition has grown significantly due to the increasing demand for accurate face detection and recognition. The widely used Haar Cascade method, however, provides suboptimal recognition and often requires image enhancement. This study aims to improve facial recognition accuracy by combining the Haar Cascade method with the Gabor Filter for image enhancement. The research was conducted in two stages: dataset training and input data testing, including experiments with variations in distance, occlusions, and camera angles. The combined method achieved optimal face detection and recognition up to 200 cm, compared to 100 cm when using Haar Cascade alone. The system successfully recognized faces partially blocked by masks, hats, goggles, or hands. Incorporating a Pan-Tilt camera further enhanced performance, achieving 99.8% accuracy at 88° pan and 84° tilt. The integration of Haar Cascade and Gabor Filter significantly improves facial recognition robustness, extending detection range, handling occlusions, and maintaining high accuracy across varying camera angles, making it suitable for advanced security applications.]]>
Copyrights © 2025
