Improving efficient face recognition with the help of hybrid optimization algorithms BRO-GOA and deep convolutional neural network
Face recognition technology, with its promising prospects, faces some limitations in the face of challenges such as posture changes, partial occlusion, expression changes, illumination, and side data. This research introduces a hybrid optimization algorithm of BRO-GOA and deep convolutional neural network to increase the accuracy and efficiency of face recognition in noisy environments. The main goal of this paper is to develop an optimization-based deep learning approach for face image recognition that is capable of processing complex and noisy data. The Battle Royale Optimization Algorithm (BRO) and Type II fuzzy system are used in this model to remove noise and improve post-processing. The results of comparative analysis show that the proposed model has the highest accuracy in all values of LFW training data and its accuracy improves from 0.9165to 0.9743with increasing data. Compared to other models, this method shows a significant reduction in the false acceptance error rate (FAR) from 0.48to 0.18and the false rejection error rate (FRR) from 0.175to 0.072. The proposed model also has a significant performance improvement at all illumination angles, especially at 0and 20degrees with accuracies of 0.9293and 0.9403. These results indicate greater stability and better performance of this model than other methods in variable conditions and noisy environments. Finally, the proposed method with high accuracy and optimal performance is a good choice for real-world applications of face recognition in complex environments.