Advancing Facial Recognition: Enhanced Model with Improved Deepface Algorithm for Robust Adaptability in Diverse Scenarios
Monali Gulhane, Sandeep Kumar, Munish Kumar, Yash Dhankhar, Bhawna Kaliraman · 2023
Face detection has been important for all the applications where security is concerned. There is a growing need to provide solutions for providing excellent and efficient protection using face detection systems. Thus, the proposed model has been the solution to give a face detection system that efficiently detects face accuracy and can see faces when multiple faces are in one image. The system can also detect faces accurately after frequent changes in the image's background. The system is robust in all the significances; if the image is scaled or compressed, then the proposed model gives higher accuracy with efficient face detection. The enhanced DeepFace algorithm has achieved an accuracy of 99.6%. The improved model is also compared with other models of facial recognition, such as OpenCV-DNN, FaceNet, Haar Cascade, and Eigenface, demonstrating respective accuracy rates of 91%, 89%, 82%, and 84.9%, indicating the improved model is better and efficient for the application of the face detection in providing security.