One of the key causes of traffic accidents globally is driver drowsiness. This demands efficient, real-time intervention methods. This paper proposes an intelligent detection method that leverages the capabilities of Convolutional Neural Networks (CNNs) in state detection and YOLOv5 in facial localization. This method also demonstrates efficient detection of the level of driver drowsiness by using biologically inspired features, which include the Eye Aspect Ratio (EAR), Mouth Aspect Ratio (MAR), and Head Pose Angle. This detection tool was trained on a diverse dataset that was comprised of different lighting sources, vehicles, and drivers. This was done in the objective of ensuring the detection method was robust. This method has a detection accuracy of 98.09% and offers a 15% reduction in training time compared to the standard CNN architecture. This method worked effectively for all ages and both genders.