Enhancing Facial Recognition Efficiency with Cloud-Based Parallel Radial Basis Function Networks

Facial recognition has high uniqueness and is difficult to forge. Facial recognition is more secure and reliable compared with other identity authentication methods, and can effectively prevent identity theft and fraud. Therefore, to ensure social and network security, a cloud computing Map-Reduce parallel optimized Radial Basis Function (RBF) neural network is built to improve the performance of facial recognition. Firstly, to optimize RBF networks, the K-means++ algorithm is taken to accurately determine the position of the hidden layer center and optimize the network structure. Secondly, to further improve the processing speed and scalability of the facial recognition system, the research also utilizes the Map-Reduce framework in cloud computing to perform parallel optimization on the RBF network. The average facial recognition accuracy was 99.6%, surpassing existing models. The model achieved a minimum recognition precision of 97.7% and an average recognition precision of 98.1%. In terms of recall rate, the lowest was 97.1% and the average was 97.7%, showing excellent performance. In addition, the F1-Score was as low as 0.982 and as high as 0.986 on average, demonstrating its efficiency in facial recognition tasks. The receiver operation characteristic curve was 0.987, further confirming its superior facial recognition ability. The above results indicate that the proposed cloud computing Map-Reduce parallel optimized RBF network has strong application potential in facial recognition, providing valuable references for future research.