Among the causes of the traffic accidents driver drowsiness comes at one of the first places. According to the literature some work has been done in driver fatigue detection. This paper proposes a Real Time Driver Fatigue Detection Based on Support Vector Machine (SVM) Algorithm. Fatigue detection mainly focuses on drivers' face expressions and behaviors. OpenCV and Dlib libraries were utilized to detect the expressions of drivers' faces. The proposed system has five stages: PERCLOS, count of yawn, internal zone of the mouth opening, count of eye blinking and head detection to extract attributes from real time video. Subsequently, facial expressions were trained with SVM. In this study an SVM-based driver fatigue detection is recommended and the tests showed that the accuracy rate of fatigue detection is up to 97.93%.

In this letter, a fast digital image stabilization method based on local binary pattern (LBP) for real time applications is presented. The LBP approach utilized in this work enables efficient representation of the image frames in 1-bit depth resolution. A simple Boolean exclusive-OR (XOR) based matching criterion is employed to decide global motion vector between consecutive image frames. The constant velocity model based Kalman filtering is executed on the global motion vectors to obtain smoothed frame positions. Experiments show that the proposed approach provides comparable or better performance against the methods at the same category by requiring lower computational complexity.

In this study, optical character recognition (OCR) was carried out by using artificial neural network. Negative correlation learning (NCL) method was used to teach artificial neural network. Negative correlation learning, which trains artificial neural networks in groups instead of teaching them individually, is a new technique. By teaching different individual networks as only one network, different parts can be taught at the same time and with this feature teaching will be better and teaching time will be shorter. In this study, image taken with mobile phone were partitioned by processing optical character recognition techniques and their properties were determined. After these letters' images was classified and converted into text by using artificial neural network ensemble which were taught by negative correlation learning respectively. This system work successfully.