Driver Alertness Monitoring Using Fusion of Facial Features and Bio-Signals

Lee, Boon-Giin; Chung, Wan-Young

doi:10.1109/jsen.2012.2190505

Cited by 148 publications

(58 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Fuzzy and neural networks [7] Needs a huge training set of face and non-face images and take a long time in execution. Eye blink pattern detection…”

Section: Existing Research Work Limitationsmentioning

confidence: 99%

Robust and Efficient Driver Monitoring System Using a Synthesis of Facial Features and Biosignals

Kumar¹,

Raj²

2015

Biosci., Biotech. Res. Asia

View full text Add to dashboard Cite

Monitoring the driver's state of consciousness and fatigue is exclusively important to reduce the number of traffic accidents. We have proposed a collective data fusion method for checking driver safety levels by combining eye features and Heart rate Variability (HRV). Fatigue behavior was determined via facial image processing using Matlab as a tool. After detecting the face, the location of the eyes will be detected using the duration of eye closure, and the frequency of eye blinks. It is used as a distinctive characteristic to judge whether a driver is drowsy or not. General duration of closure is 0.15 to 0.25 seconds. Initially the HRV analysis was accomplished by acquiring Electrocardiogram (ECG) signal through non-intrusive ECG sensors wrapped on to the steering wheel followed by filtering noises and we have calculated pulse rate via peak detection using LabVIEW. Driver fatigue and heart attacks at road are the remarkable factors for a huge number of vehicle accidents. Driver Fitness and Monitoring is responsible for reviewing driving privileges of individuals and assessing their ability to safely operate motor vehicles. Some of the diseases or disabilities that may interfere with the safe operation of vehicles can be found in these categories: Vision (e.g. minimum correct vision, double vision, telescopic lenses), Hearing (e.g. vertigo, issues particular to commercial drivers), Cardiovascular diseases (e.g. heart disease, heart attack), Cerebrovascular Diseases (e.g. stroke), Metabolic diseases (e.g. diabetes, parathyroid, pituitary, adrenal), Musculoskeletal Disabilities (e.g. arms, legs, spine, paraplegia, quadraplegia), Psychiatric disorders (e.g. mental, emotional, personality, psychotic), effects of drugs (e.g. sedatives, tranquillizers, antidepressants, narcotics), effects of Alcohol (e.g. alcohol dependency) and aging Issues (e.g. loss of strength, slow reaction time, lack of attention, confusion, progressive dementia) 19 . Among these issues, fatigue and sudden heart attack are the major causes of accidents which cannot be controlled by driver even if he/she wants to do.The post-operative patients can develop complications once they are discharged from hospital. Hence the ECG of such patients needs to be monitored for some time after their treatment.

show abstract

“…Fuzzy and neural networks [7] Needs a huge training set of face and non-face images and take a long time in execution. Eye blink pattern detection…”

Section: Existing Research Work Limitationsmentioning

confidence: 99%

Robust and Efficient Driver Monitoring System Using a Synthesis of Facial Features and Biosignals

Kumar¹,

Raj²

2015

Biosci., Biotech. Res. Asia

View full text Add to dashboard Cite

show abstract

“…Kalman filters, which are commonly used in general sensor fusion and have been shown to improve psychophysiological inference with autonomic nervous system responses by 'learning' about a subject over time (Koenig et al 2011;Novak et al 2011), are theoretically a simple dynamic Bayesian network, though not wellvalidated in physiological computing. More advanced dynamic Bayesian networks have been tested, some of them incorporating context-awareness (Ji et al 2006;Lee and Chung 2012;Yang et al 2008). Besides Bayesian networks, alternate dynamic classifiers include e.g.…”

Section: Dynamic and Ensemble Classification Algorithmsmentioning

confidence: 99%

Engineering Issues in Physiological Computing

Novak

2014

Human–Computer Interaction Series

View full text Add to dashboard Cite

Prototypes of physiological computing systems have appeared in countless fields, but few have made the leap from research to widespread use. This is due to several practical problems that can be roughly divided into four major categories: hardware, signal processing, psychophysiological inference, and feedback loop design. This chapter explores these issues from an engineering point of view, discussing major weaknesses and suggesting directions for potential solutions. Specifically, some of the topics covered are: unobtrusiveness and robustness of the hardware, real-time signal processing capability, different approaches to design and validation of a psychophysiological classifier, and the desired complexity of the feedback rules. The chapter also briefly discusses the challenge of finding an appropriate practical application for physiological computing, then ends with a summary of recommendations for future research.

show abstract

“…For example, if symptoms of fatigue, falling asleep, or fainting have been detected, an alarm signal would be triggered to stop the vehicle or device. 3 In addition, both military and civil aviation utilize systems for monitoring pilots to monitor their psychophysiological state, 4 ensuring pilot and passenger safety as well as the mission outcome, aircraft condition, and load being transported.…”

Section: Introductionmentioning

confidence: 99%

Fiber-optic sensors for monitoring patient physiological parameters: a review of applicable technologies and relevance to use during magnetic resonance imaging procedures

Dziuda

2015

J. Biomed. Opt.

View full text Add to dashboard Cite

Abstract. The issues involved with recording vital functions in the magnetic resonance imaging (MRI) environment using fiber-optic sensors are considered in this paper. Basic physiological parameters, such as respiration and heart rate, are fundamental for predicting the risk of anxiety, panic, and claustrophobic episodes in patients undergoing MRI examinations. Electronic transducers are generally hazardous to the patient and are prone to erroneous operation in heavily electromagnetically penetrated MRI environments; however, nonmetallic fiberoptic sensors are inherently immune to electromagnetic effects and will be crucial for acquiring the abovementioned physiological parameters. Forty-seven MRI-tested or potentially MRI-compatible sensors have appeared in the literature over the last 20 years. The author classifies these sensors into several categories and subcategories, depending on the sensing element placement, method of application, and measurand type. The author includes five in-house-designed fiber Bragg grating based sensors and shares experience in acquiring physiological measurements during MRI scans. This paper aims to systematize the knowledge of fiber-optic techniques for recording life functions and to indicate the current directions of development in this area.

show abstract

Driver Alertness Monitoring Using Fusion of Facial Features and Bio-Signals

Cited by 148 publications

References 12 publications

Robust and Efficient Driver Monitoring System Using a Synthesis of Facial Features and Biosignals

Robust and Efficient Driver Monitoring System Using a Synthesis of Facial Features and Biosignals

Engineering Issues in Physiological Computing

Fiber-optic sensors for monitoring patient physiological parameters: a review of applicable technologies and relevance to use during magnetic resonance imaging procedures

Contact Info

Product

Resources

About