<p>Face detection and recognition support the unique identification of individuals in the real-world environment, which is suitable for applications such as crowd control, surveillance, user authentication in mobile phones, photo tagging in social media, tracking attendance, and so on. The traditional methods in face detection and recognition particularly suffer from scalability issues, security vulnerability, pose variations, background occlusion, and high false error rates. To resolve these issues, optimization-enabled Incremental learning-based MobileNet and Convolutional neural network (ILMNetCNN-TSHO) is proposed, which detects the facial samples effectively. Additionally, integrated incremental learning improves the model's knowledge by accessing the previous tasks and securing the obtained information for achieving the desired recognition performance. Furthermore, the efficacy of the developed ILMNetCNN is improved by the integration of the Twisted stepping Hierarchical Optimization algorithm (TSHO) that boosts the recognition ability by handling the local optimization problems associated with vanishing gradient challenge and convergence. The performance of ILMNetCNN-TSHO in terms of Negative Predictive Value (NPV), False Positive Rate (FPR), True Positive Rate (TPR), accuracy, and Positive Predictive Value (PPV) is 0.97, 0.94, 0.95, 95.14%, and 0.93, specifically.</p>