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

Nopsuwanchai, Roongroj; Noguchi, Yoshihiro; Ohsuga, Mieko; Kamakura, Yoshiyuki; Inoue, Yoshio

doi:10.1109/itsc.2008.4732622

Cited by 13 publications

(7 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…If the frames continue to exhibit the drowsiness state for more than 3 seconds, an alarm is turned on. The average accuracy (AAC), the detection rate (DR) and false alarm rate (FAR) in accordance with equations (4), (5) and (6) has been calculated. These three measures, which have been proposed for evaluating the accuracy, indicate the acceptable performance of the proposed algorithm in detecting the signs of fatigue in driver's face at the time of driving.…”

Section: Results Analysismentioning

confidence: 99%

Driver’s Drowsiness Detection by Real Time Facial Features Monitoring

Teney.J.¹,

Rani²

2019

IJRTER

View full text Add to dashboard Cite

To monitor the drowsiness of driver, this paper describes an efficient method by using three well defined phases. The threephases are facial features detection using Viola Jones, the eyetracking and yawning detection. Once the face is detected, the system is made illumination invariant by segmenting the skin part alone and considering only the chromatic components to reject most of the non face image backgrounds based on skin color. The tracking of eyes and yawning detection are done by correlation coefficient template matching. They can easily capture the image of the personby using a single camera in all directions. The feature vectors from each of the above phases are concatenated and a binarylinear support vector machine classifier is used to classify the consecutive frames into fatigue and nonfatigue states and sound an alarm for the former, if it is above the threshold time. Extensive real time experiments prove that the proposed method is highly efficient in finding the drowsiness and alerting the driver.

show abstract

Section: Results Analysismentioning

confidence: 99%

Driver’s Drowsiness Detection by Real Time Facial Features Monitoring

Teney.J.¹,

Rani²

2019

IJRTER

View full text Add to dashboard Cite

show abstract

“…The scientific literature generally quantizes the ground truth of drowsiness as a discrete LoD taking N distinct integer values (with

) and annotated based on various indicators of drowsiness. The ground-truth LoD can be self-annotated by subjects in terms of a subjective questionnaire [ 14 , 15 ], marked positive when line crossings occur in a driving simulator [ 16 ] or on real roads [ 17 ], annotated by trained experts by visually looking for physiological indicators of drowsiness in the brain signals [ 17 , 18 ] or in the face video [ 19 , 20 ], or non-spontaneously acted out by subjects according to a pre-defined, given script [ 21 , 22 , 23 ].…”

Section: Background On Automatic Real-time Characterization Of Drmentioning

confidence: 99%

“…Therefore, such systems generally use machine learning models trained in a supervised manner, which requires a ground truth to be available. In practice, these learned systems typically adopt the cascade structure that consists in first (1) extracting an intermediate representation, e.g., a vector of features [ 14 , 15 , 16 , 17 , 18 , 19 ] or a sequence of features [ 20 , 21 , 22 ], and then (2) characterizing drowsiness, as defined by the selected type of ground truth. Note that these features generally consist of standard measures of objective indicators, such as the percentage of eye closure (PERCLOS) and the standard deviation of lateral position (SDLP).…”

Section: Background On Automatic Real-time Characterization Of Drmentioning

confidence: 99%

“…For extracting the intermediate representation, algorithms consist of proprietary softwares [ 14 , 17 , 18 ], face landmarks alignment [ 16 , 20 , 21 ], thresholds on the first derivative of the electro-oculogram (EOG) signal [ 15 ], adaptative image filters and statistical fitting [ 19 ], or a pre-trained CNN [ 22 ] such as the VGG-16 one [ 24 ]. For characterizing drowsiness, models consist of logistic regression [ 16 , 17 ], support vector machine (SVM) [ 15 ], artificial neural network (ANN) [ 14 , 15 , 19 ], hidden Markov model (HMM) [ 20 , 21 ], long-short term memory (LSTM) network smoothed by a temporal CNN [ 22 ], or end-to-end 3D-CNN [ 23 ]. Table 1 lists the design choices made by others in the field of automatic, real-time drowsiness characterization systems based on faces images, and compares them to the ones of our system (in bold).…”

Section: Background On Automatic Real-time Characterization Of Drmentioning

confidence: 99%

See 1 more Smart Citation

Multi-Timescale Drowsiness Characterization Based on a Video of a Driver’s Face

Massoz

Verly

Droogenbroeck

2018

Sensors

View full text Add to dashboard Cite

Drowsiness is a major cause of fatal accidents, in particular in transportation. It is therefore crucial to develop automatic, real-time drowsiness characterization systems designed to issue accurate and timely warnings of drowsiness to the driver. In practice, the least intrusive, physiology-based approach is to remotely monitor, via cameras, facial expressions indicative of drowsiness such as slow and long eye closures. Since the system’s decisions are based upon facial expressions in a given time window, there exists a trade-off between accuracy (best achieved with long windows, i.e., at long timescales) and responsiveness (best achieved with short windows, i.e., at short timescales). To deal with this trade-off, we develop a multi-timescale drowsiness characterization system composed of four binary drowsiness classifiers operating at four distinct timescales (5 s, 15 s, 30 s, and 60 s) and trained jointly. We introduce a multi-timescale ground truth of drowsiness, based on the reaction times (RTs) performed during standard Psychomotor Vigilance Tasks (PVTs), that strategically enables our system to characterize drowsiness with diverse trade-offs between accuracy and responsiveness. We evaluated our system on 29 subjects via leave-one-subject-out cross-validation and obtained strong results, i.e., global accuracies of 70%, 85%, 89%, and 94% for the four classifiers operating at increasing timescales, respectively.

show abstract

“…In addition to medical treatment and healthcare, the biosignals could also be considered as a significant interface between human body and machines. In the literature, applications like activity recognition [ 4 ], driving assistance [ 5 , 6 ] or human-computer interface [ 7 , 8 , 9 ] were mentioned by researchers. The potential of biosignals is still to be exploited when sensing technologies advance further.…”

Section: Introductionmentioning

confidence: 99%

Common-Mode Noise Reduction in Noncontact Biopotential Acquisition Circuit Based on Imbalance Cancellation of Electrode-Body Impedance

Chen

Wang

Anzai

et al. 2020

Sensors

View full text Add to dashboard Cite

Biopotential sensing technology with electrodes has a great future in medical treatment and human—machine interface, whereas comfort and longevity are two significant problems during usage. Noncontact electrode is a promising alternative to achieve more comfortable and long term biopotential signal recordings than contact electrode. However, it could pick up a significantly higher level of common-mode (CM) noise, which is hardly solved with passive filtering. The impedance imbalance at the electrode-body interface is a limiting factor of this problem, which reduces the common mode rejection ratio (CMRR) of the amplifier. In this work, we firstly present two novel CM noise reduction circuit designs. The circuit designs are based on electrode-body impedance imbalance cancellation. We perform circuit analysis and circuit simulations to explain the principles of the two circuits, both of which showed effectiveness in CM noise rejection. Secondly, we proposed a practical approach to detect and monitor the electrode-body impedance imbalance change. Compared with the conventional approach, it has certain advantages in interference immunity, and good linearity for capacitance. Lastly, we show experimental evaluation results on one of the designs we proposed. The results indicated the validity and feasibility of the approach.

show abstract

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

Cited by 13 publications

References 8 publications

Driver’s Drowsiness Detection by Real Time Facial Features Monitoring

Driver’s Drowsiness Detection by Real Time Facial Features Monitoring

Multi-Timescale Drowsiness Characterization Based on a Video of a Driver’s Face

Common-Mode Noise Reduction in Noncontact Biopotential Acquisition Circuit Based on Imbalance Cancellation of Electrode-Body Impedance

Contact Info

Product

Resources

About