Automatic Event Recognition Research Articles

Intelligent Natural Language Processing for Epidemic Intelligence

Epidemic Intelligence activities depend significantly on analysts’ ability to locate and aggregate heterogeneous and complex information promptly. The level of novelty of the targeted information is a challenge. The earlier events of interest are located the larger the benefit: more accurate and timely warnings can be made available by the analysts. In this work, the role of Natural Language Processing technologies is investigated. In particular, transformer-based encoding of Web documents (such as newspaper articles as well as epidemic bulletins) for the automatic recognition of events and relevant epidemic information is adopted and evaluated. The resulting framework is configured as a domain-specific meta-search methodology and as a possible basis for a novel generation of Web search environments supporting the Epidemic Intelligence analyst.

The concept of an automatic system for preliminary entry of hazardous facility into emergency mode with an approaching earthquake

The article discusses the concept of introducing an automatic system for preliminary input of hazardous facilities into emergency mode in the event of an approaching earthquake, developed at the South Ukrainian National Pedagogical University named after K. D. Ushinsky. Theoretical prerequisites for the study of heterogeneities, using a passive method based on the location of secondary waves reflected from the inhomogeneities under the basements and foundations of buildings of already operated building structures, were taken as the basis of the proposed system. Due to the finite speed of propagation of seismic waves and having a sensor in the focus of an earthquake, it is possible for distant objects to learn about an impending disaster in a few tens of seconds. The effective use of the time window from the focus of the earthquake to hazardous facilities requires the development of: measuring equipment, data transmission systems, automatic event recognition, elements for automating the process of putting hazardous objects into emergency mode- all this is the goal and objectives of this article.

Olympic Games Event Recognition via Transfer Learning with Photobombing Guided Data Augmentation.

Automatic event recognition in sports photos is both an interesting and valuable research topic in the field of computer vision and deep learning. With the rapid increase and the explosive spread of data, which is being captured momentarily, the need for fast and precise access to the right information has become a challenging task with considerable importance for multiple practical applications, i.e., sports image and video search, sport data analysis, healthcare monitoring applications, monitoring and surveillance systems for indoor and outdoor activities, and video captioning. In this paper, we evaluate different deep learning models in recognizing and interpreting the sport events in the Olympic Games. To this end, we collect a dataset dubbed Olympic Games Event Image Dataset (OGED) including 10 different sport events scheduled for the Olympic Games Tokyo 2020. Then, the transfer learning is applied on three popular deep convolutional neural network architectures, namely, AlexNet, VGG-16 and ResNet-50 along with various data augmentation methods. Extensive experiments show that ResNet-50 with the proposed photobombing guided data augmentation achieves 90% in terms of accuracy.

Analytical transfer function for the nonlinear response of a resonant medium in the spatio-temporal Fourier-transform domain

The nonlinear response of a resonant medium has many applications. To model and find the response of such a medium requires solving the Schrodinger equation (SE), which is a computationally extensive task. In this paper, we develop an analytical model to find the response of a resonant medium due to an applied field by employing the spatio-temporal Fourier-transform (STFT)–domain-based transfer function. A key feature of this approach is the use of the resonant excitation approximation (REA), which amounts to assuming that a group of atoms (or other quantum systems) within a volume element in the STFT domain are excited by only the corresponding volume element in the STFT domain of the field. We first derive the one-dimensional transfer function using an inhomogeneously broadened atomic medium under the REA. Then, we develop the three-dimensional transfer function and show that the analytical model agrees closely with the results obtained via an explicit simulation of the atomic response. As a practical example of the analytical model, we show that it can be used to model a spatio-temporal-correlator-based automatic event recognition system at a speed that is many orders of magnitude faster than solving the SE.

All optical three dimensional spatio-temporal correlator for automatic event recognition using a multiphoton atomic system

We describe an automatic event recognition (AER) system based on a three-dimensional spatio-temporal correlator (STC) that combines the techniques of holographic correlation and photon echo based temporal pattern recognition. The STC is shift invariant in space and time. It can be used to recognize rapidly an event (e.g., a short video clip) that may be present in a large video file, and determine the temporal location of the event. Using polar Mellin transform, it is possible to realize an STC that is also scale and rotation invariant spatially. Numerical simulation results of such a system are presented using quantum mechanical equations of evolution. For this simulation we have used the model of an idealized, decay-free two level system of atoms with an inhomogeneous broadening that is larger than the inverse of the temporal resolution of the data stream. We show how such a system can be realized by using a lambda-type three level system in atomic vapor, via adiabatic elimination of the intermediate state. We have also developed analytically a three dimensional transfer function of the system, and shown that it agrees closely with the results obtained via explicit simulation of the atomic response. The analytical transfer function can be used to determine the response of an STC very rapidly. In addition to the correlation signal, other nonlinear terms appear in the explicit numerical model. These terms are also verified by the analytical model. We describe how the AER can be operated in a manner such that the correlation signal remains unaffected by the additional nonlinear terms. We also show how such a practical STC can be realized using a combination of a porous-glass based Rb vapor cell, a holographic video disc, and a lithium niobate crystal.

Assessment of Autonomy in Instrumental Activities of Daily Living in Pre-and Demented Patients Using an Automatic Video Monitoring System

Objectives To investigate the use of a video monitoring system for automatic event recognition for the assessment of autonomy in Instrumental Activities of Daily Living (IADL) in dementia patients. Methods Three groups of participants (healthy control, Mild Cognitive Impairment and Alzheimer's disease) had to carry out a standardized scenario consisting of directed tasks (single and dual task) and IADLs such as preparing pillbox. During this time they were recorded by 3D video cameras capturing all their activities. The performance quality of each participant was manually annotated and assessed based on the amount of successfully carried out activities. Recorded data was processed by a platform of video signal analysis in order to extract kinematic parameters detecting activities undertaken by the participant. We developed a classifier based on the extracted video features for diagnostic prediction and further autonomy performance prediction. Results Overall activities were correctly automatically detected. The most accurate detected activities were: using the phone with 91% accuracy and preparing pillbox with 88% accuracy. The diagnostic group classifier based on the automatically extracted video features obtained accuracy of 71.79 % when combining directed tasks and IADLs. Autonomy group classifier obtained an accuracy of 84.61% when combining directed tasks and IADLs. Conclusions The results suggest that it is possible to assess autonomy with the help automatic video monitoring system (AVMS) and that the use of such technologies could provide clinicians with diagnostic relevant information and improve autonomy assessment in real time decreasing observer biases.

Validation of an automatic video monitoring system for the detection of instrumental activities of daily living in dementia patients.

Over the last few years, the use of new technologies for the support of elderly people and in particular dementia patients received increasing interest. We investigated the use of a video monitoring system for automatic event recognition for the assessment of instrumental activities of daily living (IADL) in dementia patients. Participants (19 healthy subjects (HC) and 19 mild cognitive impairment (MCI) patients) had to carry out a standardized scenario consisting of several IADLs such as making a phone call while they were recorded by 2D video cameras. After the recording session, data was processed by a platform of video signal analysis in order to extract kinematic parameters detecting activities undertaken by the participant. We compared our automated activity quality prediction as well as cognitive health prediction with direct observation annotation and neuropsychological assessment scores. With a sensitivity of 85.31% and a precision of 75.90%, the overall activities were correctly automatically detected. Activity frequency differed significantly between MCI and HC participants (p < 0.05). In all activities, differences in the execution time could be identified in the manually and automatically extracted data. We obtained statistically significant correlations between manually as automatically extracted parameters and neuropsychological test scores (p < 0.05). However, no significant differences were found between the groups according to the IADL scale. The results suggest that it is possible to assess IADL functioning with the help of an automatic video monitoring system and that even based on the extracted data, significant group differences can be obtained.

Real-time change detection in data streams with FPGAs

The automatic recognition of changes in data streams is useful in both real-time and off-line data analyses. This article shows several effective change-detecting algorithms (based on martingales) and describes their real-time applicability in the data acquisition systems through the use of Field Programmable Gate Arrays (FPGA). The automatic event recognition system is absolutely general and it does not depend on either the particular event to detect or the specific data representation (waveforms, images or multidimensional signals). The developed approach provides good results for change detection in both the temporal evolution of profiles and the two-dimensional spatial distribution of volume emission intensity. The average computation time in the FPGA is 210μs per profile.

A flexible framework for recognizing events, temporal expressions, and temporal relations in clinical text

To provide a natural language processing method for the automatic recognition of events, temporal expressions, and temporal relations in clinical records. A combination of supervised, unsupervised, and rule-based methods were used. Supervised methods include conditional random fields and support vector machines. A flexible automated feature selection technique was used to select the best subset of features for each supervised task. Unsupervised methods include Brown clustering on several corpora, which result in our method being considered semisupervised. On the 2012 Informatics for Integrating Biology and the Bedside (i2b2) shared task data, we achieved an overall event F1-measure of 0.8045, an overall temporal expression F1-measure of 0.6154, an overall temporal link detection F1-measure of 0.5594, and an end-to-end temporal link detection F1-measure of 0.5258. The most competitive system was our event recognition method, which ranked third out of the 14 participants in the event task. Analysis reveals the event recognition method has difficulty determining which modifiers to include/exclude in the event span. The temporal expression recognition method requires significantly more normalization rules, although many of these rules apply only to a small number of cases. Finally, the temporal relation recognition method requires more advanced medical knowledge and could be improved by separating the single discourse relation classifier into multiple, more targeted component classifiers. Recognizing events and temporal expressions can be achieved accurately by combining supervised and unsupervised methods, even when only minimal medical knowledge is available. Temporal normalization and temporal relation recognition, however, are far more dependent on the modeling of medical knowledge.

Semantic Analysis for Automatic Event Recognition and Segmentation of Wedding Ceremony Videos

Wedding is one of the most important ceremonies in our lives. It symbolizes the birth and creation of a new family. In this paper, we present a system for automatically segmenting a wedding ceremony video into a sequence of recognizable wedding events, e.g., the couple's wedding kiss. Our goal is to develop an automatic tool that helps users to efficiently organize, search, and retrieve his/her treasured wedding memories. Furthermore, the obtained event descriptions could benefit and complement the current research in semantic video understanding. Based on the knowledge of wedding customs, a set of audiovisual features, relating to the wedding contexts of speech/music types, applause activities, picture-taking activities, and leading roles, are exploited to build statistical models for each wedding event. Thirteen wedding events are then recognized by a hidden Markov model, which takes into account both the fitness of observed features and the temporal rationality of event ordering to improve the segmentation accuracy. We conducted experiments on a collection of wedding videos and the promising results demonstrate the effectiveness of our approach. Comparisons with conditional random fields show that the proposed approach is more effective in this application domain.

INVESTIGATION OF REACTOR CONDITION MONITORING AND SINGULARITY DETECTION VIA WAVELET TRANSFORM AND DE-NOISING

Wavelet theory was applied to detect a singularity in a reactor power signal. Compared to Fourier transform, wavelet transform has localization properties in space and frequency. Therefore, using wavelet transform after de-noising, singular points can easily be found. To test this theory, reactor power signals were generated using the HANARO(a Korean multi-purpose research reactor) dynamics model consisting of 39 nonlinear differential equations contaminated with Gaussian noise. Wavelet transform decomposition and de-noising procedures were applied to these signals. It was possible to detect singular events such as a sudden reactivity change and abrupt intrinsic property changes. Thus, this method could be profitably utilized in a real-time system for automatic event recognition(e.g., reactor condition monitoring).

A portable microsystem-based telemetric pressure and temperature measurement unit.

This article shows how a fully implantable stand-by device for measuring intracorporal pressure and temperature under normal conditions can be implemented, consisting of a sensor element combined with a transcutaneous telemetric interface. One further point of interest is automatic event recognition in order to capture special signal components in an emergency situation. Therefore, signal processing and waveform analysis are exigent, first to observe the measured signal in realtime on a portable unit, and second to process the data offline on a stationary unit.

Automatic detection of esophageal pressure events

Ambulatory long-term motility recording is used increasingly for evaluation of esophageal function. The enormous amount of motility data recorded by this method demands subsequent computer analysis. One of the most crucial steps of this analysis becomes the process of automatic selection of relevant pressure peaks at the various recording levels. Until now, this selection has been performed entirely by rule-based systems, requiring each pressure deflection to fit within predefined rigid numerical limits in order to be detected. However, due to great variations in the shapes of the pressure curves generated by muscular contractions, rule-based criteria do not always select the pressure events most relevant for further analysis. We have therefore been searching for a new concept for automatic event recognition. The present study describes a new system, based on the method of neurocomputing. A large sample of normal esophageal pressure deflections was used as a "learning set," and the performance of the trained neural networks was subsequently verified on different sets of data from normal subjects. Our trained networks detected pressure deflections with sensitivities of 0.79-0.99 and accuracies of 0.89-0.98, depending on the recording level within the esophageal lumen. The neural networks often recognized peaks that clearly represented true contractions but that had been rejected by a rule-based system. We conclude that neural networks have potentials for automatic detections of esophageal, and possibly also other kinds of gastrointestinal, pressure variations.

Application des réseaux de neurones pour l'analyse de l'articulation musicale

Knowledge discovery approach was made in the domain of pipe organ sounds. The background of such an approach was reviewed and some experimental results were quoted. Parameterization of pipe organ sounds was discussed, and a feature extraction method was described in terms of parameters derived from both time- and frequency-domain analyses. The knowledge acquisition system principles were described and the neural network algorithms underlying both training and performance processes were presented. The number of necessary data inputs and the structure of neural net were studied in order to obtain the demanded accuracy of results. Experimental results have shown that the technique applied to the analysis of musical events in pipe organ sound might form an approach to the domain of automatic recognition of events in sound