No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Оптимизация искусственных нейронных сетей в задачах обработки графической информации для идентификации психофизиологических состояний субъекта
Abstract
Использование дистанционных технологий идентификации психофизиологических состояний (ПФС) субъекта необходимо. Применение таких систем имеет ряд преимуществ, связанных с возможностью осуществления скрытого контроля, с отсутствием физического контакта человека с системой и т. д. В связи с распространением коронавирусной инфекции COVID-19 индустрия безопасности нахо-дится в поиске способов использования имеющихся решений, в частности на базе тепловизионных ка-мер, для их интеграции в системы массового скрининга субъектов. Это сделало очевидным тот факт, что тепловидение является альтернативным инструментом в борьбе с распространением эпидемии. Современные системы оценки ПФС человека имеют либо недостаточный функционал, связанный с ограниченным кругом идентифицируемых состояний, либо недостаточную точность идентификации состояний. Комплексирование различных методов обработки и преобразования цифровых изображений (термограмм), а также методов принятия решений на базе статистических и нейросетевых алгоритмов может решить указанную проблему. В настоящей работе приведены результаты исследований по идентификации нескольких психофизиологических состояний с использованием методов и алгоритмов обработки цифровых изображений и нейросетевого алгоритма принятия решений на базе комитета обученных сверточных нейронных сетей.
ResearchGate has not been able to resolve any citations for this publication.
Background and Objective: Driver drowsiness is a cause of many traffic accidents all around the world. Driver respiration dynamics undergoes significant changes from wakefulness to sleep. The intrusive nature of current respiration monitoring methods makes them unattractive for detection of in-vehicle driver drowsiness. In this paper, changes in the respiration rate were monitored for drowsiness detection using thermal imaging, which is completely contact-free and non-intrusive.
Materials and Methods: For each frame, the driver’s face was isolated from the rest of the image. Then, the nostrils' zone was localized using physiological characteristics of face. The respiration signal was constructed by putting together the mean temperature of nostril region in all of the frames. In order to study respiration variations from wakefulness to drowsiness, a total number of 12 subjects were tested in a driving simulator. The observer rating of drowsiness (ORD) method was used to estimate the drowsiness level of the subjects.
Results: Derivation of driver respiration rate using thermal imaging was a reliable and non-intrusive method. The results were strongly correlated with those of traditional methods. Driver respiration rate decreased from wakefulness to extreme drowsiness by 3.5 breaths per minute (bpm) and its standard deviation (SD) increased by 0.7 bpm.
Conclusion: Driver respiration rate decreased and its SD increased from wakefulness to drowsiness in 12 participants of this study. The results showed that driver drowsiness could be detected even at moderate levels from analysis of the respiration rate.
.
The aim of this study is to propose a system that is capable of recognising the identity of a person, indicating whether the person is drunk using only information extracted from thermal face images. The proposed system is divided into two stages, face recognition and classification. In the face recognition stage, test images are recognised using robust face recognition algorithms: Weber local descriptor (WLD) and local binary pattern (LBP). The classification stage uses Fisher linear discriminant to reduce the dimensionality of the features, and those features are classified using a classifier based on a Gaussian mixture model, creating a classification space for each person, extending the state-of-the-art concept of a “DrunkSpace Classifier.” The system was validated using a new drunk person database, which was specially designed for this work. The main results show that the performance of the face recognition stage was 100% with both algorithms, while the drunk identification saw a performance of 86.96%, which is a very promising result considering 46 individuals for our database in comparison with others that can be found in the literature.
To achieve the simultaneous and unobtrusive breathing rate (BR) and heart rate (HR) measurements during nighttime, we leverage a far-infrared imager and an infrared camera equipped with IR-Cut lens and an infrared lighting array to develop a dual-camera imaging system. A custom-built cascade face classifier, containing the conventional Adaboost model and fully convolutional network trained by 32K images, was used to detect the face region in registered infrared images. The region of interest (ROI) inclusive of mouth and nose regions was afterwards confirmed by the discriminative regression and coordinate conversions of three selected landmarks. Subsequently, a tracking algorithm based on spatio-temporal context learning was applied for following the ROI in thermal video, and the raw signal was synchronously extracted. Finally, a custom-made time-domain signal analysis approach was developed for the determinations of BR and HR. A dual-mode sleep video database, including the videos obtained under environment where illumination intensity ranged from 0 to 3 Lux, was constructed to evaluate the effectiveness of the proposed system and algorithms. In linear regression analysis, the determination coefficient (R²) of 0.831 had been observed for the measured BR and reference BR, and this value was 0.933 for HR measurement. In addition, the Bland-Altman plots of BR and HR demonstrated that almost all the data points located within their own 95% limits of agreement. Consequently, the overall performance of the proposed technique is acceptable for BR and HR estimations during nighttime.
In this work, facial thermal infrared images are employed for intoxicated person discrimination. Specifically, the region of the forehead of the face of the sober and the corresponding intoxicated person is used to test if the employed Local Difference Patterns (LDPs) constitute discriminative features. For an intoxicated person, vessels on the forehead become more active so that the intensity of the pixels in this region is affected accordingly. The LDPs employed ignore orientation of the pixels distribution and give emphasis on the first and second norms of the differences as well as the ordered values of the pixels in the employed kernels. The statistics of the LDPs for the drunk person are different from those of the sober one and accordingly drunkenness can be ascertained by comparing the thermal infrared image of the corresponding sober and intoxicated person. Six from the eight LDPs examined to be used as features for drunk identification were proved successful. Their classification success rate was over 73 and up to 85%. The proposed method can be incorporated into a non-invasive inspection commercial system to be used by the police as a first step for intoxicated person detection. Forty one participants in the experiment have contributed to the creation of the unique sober–drunk database which is available on the web and contains over 4.000 images.
Stress is a serious concern facing our world today, motivating the development of a better objective understanding through the use of non-intrusive means for stress recognition by reducing restrictions to natural human behavior. As an initial step in computer vision-based stress detection, this paper proposes a temporal thermal spectrum (TS) and visible spectrum (VS) video database ANUStressDB -a major contribution to stress research. The database contains videos of 35 subjects watching stressed and not-stressed film clips validated by the subjects. We present the experiment and the process conducted to acquire videos of subjects' faces while they watched the films for the ANUStressDB. Further, a baseline model based on computing local binary patterns on three orthogonal planes (LBP-TOP) descriptor on VS and TS videos for stress detection is presented. A LBP-TOP-inspired descriptor was used to capture dynamic thermal patterns in histograms (HDTP) which exploited spatio-temporal characteristics in TS videos. Support vector machines were used for our stress detection model. A genetic algorithm was used to select salient facial block divisions for stress classification and to determine whether certain regions of the face of subjects showed better stress patterns. Results showed that a fusion of facial patterns from VS and TS videos produced statistically significantly better stress recognition rates than patterns from VS or TS videos used in isolation. Moreover, the genetic algorithm selection method led to statistically significantly better stress detection rates than classifiers that used all the facial block divisions. In addition, the best stress recognition rate was obtained from HDTP features fused with LBP-TOP features for TS and VS videos using a hybrid of a genetic algorithm and a support vector machine stress detection model. The model produced an accuracy of 86%.
High resolution thermal infrared imaging is a pioneering method giving indices of sympathetic activity via the contact-free recording of facial tissues (thermal imprints). Compared to established stress markers, the great advantage of this method is its non-invasiveness. The goal of our study was to pilot the use of thermal infrared imaging in the classical setting of human stress research. Thermal imprints were compared to established stress markers (heart rate, heart rate variability, finger temperature, alpha-amylase and cortisol) in 15 participants undergoing anticipation, stress and recovery phases of two laboratory stress tests, the Cold Pressor Test and the Trier Social Stress Test. The majority of the thermal imprints proved to be change-sensitive in both tests. While correlations between the thermal imprints and established stress markers were mostly non-significant, the thermal imprints (but not the established stress makers) did correlate with stress-induced mood changes. Multivariate pattern analysis revealed that in contrast to the established stress markers the thermal imprints could not disambiguate anticipation, stress and recovery phases of both tests. Overall, these results suggest that thermal infrared imaging is a valuable method for the estimation of sympathetic activity in the stress laboratory setting. The use of this non-invasive method may be particularly beneficial for covert recordings, in the study of special populations showing difficulties in complying with the standard instruments of data collection and in the domain of psychophysiological covariance research. Meanwhile, the established stress markers seem to be superior when it comes to the characterization of complex physiological states during the different phases of the stress cycle.
Driving Drunk or High Puts Everyone in Danger
- Greenbauer Lynn
lynn.greenbauer.ctr@dot.gov. Driving Drunk or High Puts
Everyone in Danger [Электронный ресурс]: Text // NHTSA.
2017. URL: https://www.nhtsa.gov/drunk-driving/drive-sober-orget-pulled-over (дата обращения: 08.10.2019).