Yoshiyuki Kamakura

Driver-Independent Assessment of Arousal States from Video Sequences Based on the Classification of Eyeblink Patterns

In this paper, we propose a novel approach to assess drivers arousal states based on the analysis of eyeblink characteristics. We focus on a non-intrusive and driver-independent system. We use Hidden Markov Models (HMMs) to classify eyeblink patterns from the video of the drivers, and the arousal states are estimated from the histogram variations of these typical blink patterns. A strong correlation between the eyeblink patterns derived from this approach and those derived from the recorded EOG (electro-occulography) waveforms can be observed. The arousal assessment results are also verified against the rating results by a trained rater.

Classification of blink waveforms toward the assessment of driver's arousal levels - an EOG approach and the correlation with physiological measures

The goal of this research is to develop a method to assess the arousal states using facial images of drivers. This paper refers the preparatory study on the classification of blink waveforms obtained from electo-oculogram. The transitions of the distribution of classified blinks during a simulated driving task were studied for around fifty volunteers of both genders and a wide range of generations. It was shown that the blink class ratio supposed to be under the influence of not only the subjects drowsiness levels but also by his/her behavior to battle with drowsiness. The correlation with multidimensional physiological indices was also discussed.

Classification of blink waveforms towards the assessment of driver's arousal level - an approach for HMM based classification from blinking video sequence

With recent advances in image recognition, the assessment of drivers arousal level using blinking image sequences has been expected. In this paper, we demonstrated the possibility of assessing drivers arousal level by analyzing blinking image sequences. We focused on some typical blink waveform patterns occurred under drowsy condition. We used the results of EOG (Electro-occulogram) waveform clustering as the baseline for HMM (Hidden Markov Model) blinking labeling due to the difficulty of defining blinking labels from blinking image sequence. The blink pattern classes were classified by using the HMMs based on blinking image sequences. The drivers arousal level was finally estimated by histogram variation per minute of those typical blink classes.

The Assessment of Driver's Arousal States from the Classification of Eye-Blink Patterns

To realize the real-time assessment of drivers arousal states, we propose the assessment method based on the analysis of eye-blink characteristics form image sequences. The drivers arousal level while driving is not monotonous falling from high to low. We proposed the two-dimensional arousal states transition model which was taken into account the fact that a driver usually held out against sleepiness. The eye-blink pattern categories were classified from image sequence using HMM (Hidden Markov Model), then the drivers arousal states were finally assessed using HMM by histogram distribution of those typical eye-blink categories. The arousal assessment results are also verified against the rating results by trained raters.