A Study On Face Detection Using HAAR CASCADE and YOLOV5

Abstract

Face recognition has emerged as a prominent technology with numerous applications in various fields, such as surveillance, security systems, and human-computer interaction. This thesis presents a comprehensive study on the development of a robust face recognition system using OpenCV, Python, and YOLOv5. The primary objective of this research is to design and implement an accurate and efficient face recognition system that can reliably detect and identify individuals in real-time. The proposed methodology leverages the power of OpenCV, a widely-used computer vision library, and YOLOv5, a state-of-the-art object detection algorithm, to achieve superior performance. The thesis begins with an in-depth literature review, exploring the existing approaches and methodologies in face recognition. This review serves as the foundation for the subsequent research and guides the selection and implementation of the proposed system. The research work focuses on three main stages: face detection, feature extraction, and face matching. The face detection phase employs the YOLOv5 algorithm, which utilizes deep learning techniques to accurately locate faces in an image or video stream. The detected faces are then subjected to feature extraction using OpenCV, which extracts discriminative facial features from the detected regions. To achieve robust face recognition, the extracted facial features are compared against a pre-existing database of known individuals. The thesis explores various methods for feature comparison, such as eigenfaces, Fisherfaces, and deep learning-based approaches. Experimental evaluations are conducted to analyze the performance of each method and identify the most effective approach for our system. The developed face recognition system is evaluated using extensive datasets and performance metrics, including accuracy, precision, recall, and execution time. Comparisons are made with existing face recognition systems to assess the proposed system's efficiency and effectiveness. The results demonstrate that the proposed system achieves high accuracy and real-time performance in face detection, feature extraction, and face matching tasks. The system's robustness is also evaluated by considering various challenging scenarios, such as variations in lighting conditions, occlusions, and pose variations. In conclusion, this thesis presents a comprehensive study on the development of a face recognition system using OpenCV, Python, and YOLOv5. The research work contributes to the advancement of face recognition technology and offers valuable insights into the practical implementation of such systems. The findings provide a solid foundation for future research and development in the field of computer vision and biometrics.