In recent years, the rise of car accident fatalities has grown significantly around the world. Hence, road security has become a global concern and a challenging problem that needs to be solved. The deaths caused by road accidents are still increasing and currently viewed as a significant general medical issue. The most recent developments have made in advancing knowledge and scientific capacities of vehicles, enabling them to see and examine street situations to counteract mishaps and secure travelers. Therefore, the analysis of driver’s behaviors on the road has become one of the leading research subjects in recent years, particularly drowsiness, as it grants the most elevated factor of mishaps and is the primary source of death on roads. This paper presents a way to analyze and anticipate driver drowsiness by applying a Recurrent Neural Network over a sequence frame driver’s face. We used a dataset to shape and approve our model and implemented repetitive neural network architecture multi-layer model-based 3D Convolutional Networks to detect driver drowsiness. After a training session, we obtained a promising accuracy that approaches a 92% acceptance rate, which made it possible to develop a real-time driver monitoring system to reduce road accidents.

Over the last decade, large companies have heavily invested in the collection, storage and the extraction of structured and unstructured data related to their clients. This trend has lead to a rise of a large data volumes considered as mine of knowledge. With the appearance of more powerful computers and the application of the graphics processing unit (GPU) computing, Deep learning has appeared as a practical solution for large data sets processing and offers an alternative strategy to learning characteristics.In this paper, we consider learning a collection of high-level facial representation characteristics, referred to as Deep hidden face characteristics, through deep learning. We present a deep learning model (DeepDetect) trained on the Annotated Facial Landmarks in the Wild(AFLW) dataset witch offers large visual variations that allows learning very discriminative face representation. We compare the performance of our DeepDetectmodel with different variants of Boosting LBPfeatures, and we show that DeepDetectoutperforms the hand-crafted features based approach. A detection accuracy of 98.17% is achieved on the fusion of Feretdata-set and negative images extracted from the wild. The depiction of an picture is no longer a heavy hand-engineered operation with our suggested strategy, but a learning operation that leads to an effective solution to accurately differentiate faces from backgrounds.