The increasing use of technology by minors, along with access to various information that is not within their age, urges the need to implement effective systems to protect them from accessing inappropriate material online. The current use of parental control and content filtration is insufficient, highlighting the demand for accurate and real-time age discernment. This paper addresses the critical challenge of minor protection in digital spaces by automating age classification systems. Further optimization of the Convolutional Neural Networks (CNN) is performed with a variety of different architectural modifications and hyperparameter tuning strategies. The proposed model demonstrates superior performance compared to alternative approaches, contributing not only to improved age classification but also to strengthening online content filtering for enhanced safety measures tailored to minors.