The driver monitoring system, also known as driver attention monitor, is a vehicle safety system to assess the driver's alertness and warn the driver if needed and eventually apply the brakes. The system uses infrared sensors to monitor driver attentiveness. Specifically, the driver monitoring system includes a camera placed on the steering column which tracks the face. If the driver is not paying attention to the road ahead and a dangerous situation is detected, the system will warn the driver by flashing lights, warning sounds. If no action is taken, the vehicle will apply the brakes (a warning alarm will sound followed by a brief automatic application of the braking system).
Using deep convolutional neural networks and machine learning, the camera monitors the eyes and driver behavior to make a judgment if the driver is sleepy or distracted. It first use Open CL to construct an image and with the use of image pyramid, it defines the face detection using the database of faces in the cloud. With deeper analysis of driver's face for eye localization - eye open and close can be determined. Images captured are of VGA resolution with loseless JPEG compression - it is stored in the device memory if there is no connectivity and on resume of network it is sent to the cloud on FIFO basis along with the complete tracking of the vehicle.
The device is fitted on windshield or on the corner panel of the bus in way that the camera looks at the driver face without obstructing his view.