Event Detection

Visual surveillance is an active research topic in image processing. Transit systems are actively seeking new or improved ways to use technology to deter and respond to accidents, crime, suspicious activities, terrorism, and vandalism. Human behavior-recognition algorithms can be used proactively for prevention of incidents or reactively for investigation after the fact. This paper describes the current state-of-the-art image-processing methods for automatic-behavior-recognition techniques, with focus on the surveillance of human activities in the context of transit applications. The main goal of this survey is to provide researchers in the field with a summary of progress achieved to date and to help identify areas where further research is needed. This paper provides a thorough description of the research on relevant human behavior-recognition methods for transit surveillance. Recognition methods include single person (e.g., loitering), multipleperson interactions (e.g., fighting and personal attacks), personvehicle interactions (e.g., vehicle vandalism), and person-facility/ location interactions (e.g., object left behind and trespassing). A list of relevant behavior-recognition papers is presented, including behaviors, data sets, implementation details, and results. In addition, algorithm's weaknesses, potential research directions, and contrast with commercial capabilities as advertised by manufacturers are discussed. This paper also provides a summary of literature surveys and developments of the core technologies (i.e., low-level processing techniques) used in visual surveillance systems, including motion detection, classification of moving objects, and tracking.

Wireless Sensor Networks (WSN) are used in variety of fields which includes military, healthcare, environmental, biological, home and other commercial applications. With the huge advancement in the field of embedded computer and sensor technology, Wireless Sensor Networks (WSN), which is composed of several thousands of sensor nodes which are capable of sensing, actuating, and relaying the collected information, have made remarkable impact everywhere. This paper presents an overview of the various research issues in WSN based applications.

In this work, we address the problem of providing fast and on-line households appliance load detection in a non-intrusive way from aggregate electric energy consumption data. Enabling on-line load detection is a relevant research problem as it can unlock new grid services such as demand-side management and raises interactivity in energy awareness possibly leading to more green behaviours. To this purpose, we propose an On-line-NILM (Non-Intrusive Load Monitoring) machine learning algorithm combining two methodologies: i) Unsupervised event-based profiling and ii) Markov chain appliance load modelling. The event-based part performs event detection through contiguous and transient data segments, events clustering and matching. The resulting features are used to build household-specific appliance models from generic appliance models. Disaggregation is then performed on-line using an Additive Factorial Hidden Markov Model from the generated appliance model parameters. Our solution is implemented on the cloud and tested with public benchmark datasets. Accuracy results are presented and compared with literature solutions, showing that the proposed solution achieves on-line detection with comparable detection performance with respect to non on-line approaches.

T he rising popularity of photosharing applications on the Web has led to the generation of huge amounts of personal image collections. Browsing through image collections of such magnitude is currently supported by the use of tags. However, tags suffer from several limitations-such as polysemy, lack of uniformity, and spam-thus not presenting an adequate solution to the problem of content organization. Therefore, automated contentorganization methods are of particular importance to improve the content-consumption experience. Because it's common for users to associate their photo-captured experiences with some landmarks-for example, a tourist site or an event, such as a music concert or a gathering with friends-we can view landmarks and events as natural units of organization for large image collections. It's for this reason that automating the process of detecting such concepts in large image sets can enhance the experience of accessing massive amounts of pictorial content.

Motivation: Football is being the most popular game throughout the world because of its excitement and enjoyment, but this popular event has got some problems like decision making during the game. As the events like goals, offside, corners occur within a second, it is difficult for the referees to detect an event in a game as they have to take a decision within seconds. Bad decisions also can lead to bad quarrels and can make the most exciting game turned into a fighting game which is not expected by the supporters and the spectators. Bad decisions also can lead to quarrels among the supporters and the world has seen these types of violence a lot of times. Bad scenes are created during the matches for these bad decisions and these bad scenes make the game worse. It could be great if these decisions could be taken automatically by any determined system. Then the most fan-based game would be the cleanest game in the world as well. Many people have done a lot of research for the betterment of this game. They tried to work with a specific event and some of them worked with goal events, some of them worked with offside events and some tried to work with other events of a football match. I personally am passionate about this game and I always wanted to watch a game with no controversial decision throughout the whole game. I want to work in this field so that I can contribute to making this game as a disciplined and clean game. As offside is one of the worst and most controversial decision prone events of a game I want to research on this part of this game. I think doing research in this event will be very much challenging because of its complexity and the ball and players are always moving. Many researchers have tried to do research with this event but they could not give accuracy and could not be effective as well.

A group of 32 healthy men (M) divided into three dierent age groups, i.e. M20 years [mean 21 (SD 1); n 12], M40 [mean 40 (SD 2); n 10] and M70 [mean 71 (SD 5); n 10] volunteered as subjects for examination of maximal and explosive force production of leg extensor muscles in both isometric and dynamic actions (squat jump, SJ and counter movement jump, CMJ, and standing long-jump, SLJ). The balance test was performed on a force platform in both isometric and dynamic actions. Maximal bilateral isometric force value in M70 was lower (P < 0.001) than in M40 and as much as 46% lower (P < 0.001) than that recorded in M20 (P < 0.001). The maximal rate of force development (RFD) on the force-time curve was in M70 lower (P < 0.001) than in M40 and as much as 64% lower than in M20. The heights in SJ and CMJ and the distance in SLJ in M70 were lower (P < 0.001) than in M40 and M20 (P < 0.001). In response to modi®cations of the visual surroundings the older subjects were 24%±47% (P < 0.05 and P < 0.001) slower in their response time in reaching the lit centre (TT) and remained 20%±34% (P < 0.001) less time inside the centre (TC) from the overall time of lighting than M40 and M20, respectively. In both older groups the individual values of isometric RFD correlated signi®cantly (P < 0.05) with the individual balance values of TT and TC. The present results would suggest that the capacity for explosive force production declines drastically with increasing age, even more than maximal muscle strength. Aging may also lead to impaired balance with a decrease in event detection and speed of postural adjustments. The decreased ability to develop force rapidly in older people seems to be associated with a lower capacity for neuromuscular response in controlling postural sway.

Real-time multiple event analysis is important for reliable situational awareness and secure operation of the power system. Multiple sequential events can induce complex superimposed pattern in the data and are challenging to analyze in real time. This paper proposes a method for accurate detection, temporal localization, and classification of multiple events in real time using synchrophasor data. For detection and temporal localization, a Teager-Kaiser energy operator (TKEO) based method is proposed. For event classification, a time series classification based method using energy similarity measure (ESM) is proposed. The proposed method is tested for simulated multiple event cases in the IEEE-118 bus system using DigSilent/PowerFactory and real PMU data for the Indian grid. Index Terms-Energy similarity measure, event classification , event detection, multiple events, Teager-Kaiser energy operator (TKEO), temporal localization.

ited resources e.g. power consumption, memory and processing power. We underline the major problems of the existing paradigms for complex event detection (based on e.g. logic programming and Semantic Web), with a special focus on the major challenges which reduce the performance of real-time event detection. Finally, we demonstrate the high performance of ASP compared to that of Semantic Web.

This paper proposes an automatic event detection technique for camera anomaly by image analysis, in order to confirm good image quality and correct field of view of surveillance videos. The technique first extracts reduced-reference features from multiple regions in the surveillance image, and then detects anomaly events by analyzing variation of features when image quality decreases and field of view changes. Event detection is achieved by statistically calculating accumulated variations along temporal domain. False alarms occurred due to noise are further reduced by an online Kalman filter that can recursively smooth the features. Experiments are conducted on a set of recorded videos simulating various challenging situations. Compared with an existing method, experimental results demonstrate that our method has high precision and low false alarm rate with low time complexity.

Recently, wireless sensor networks (WSNs) have become mature enough to go beyond being simple fine-grained continuous monitoring platforms and become one of the enabling technologies for disaster early-warning systems. Event detection functionality of WSNs can be of great help and importance for (near) real-time detection of, for example, meteorological natural hazards and wild and residential fires. From the data-mining perspective, many real world events exhibit specific patterns, which can be detected by applying machine learning (ML) techniques. In this paper, we introduce ML techniques for distributed event detection in WSNs and evaluate their performance and applicability for early detection of disasters, specifically residential fires. To this end, we present a distributed event detection approach incorporating a novel reputation-based voting and the decision tree and evaluate its performance in terms of detection accuracy and time complexity.

This paper describes an audio event detection system which automatically classifies an audio event as ambient noise, scream or gunshot. The classification system uses two parallel GMM classifiers for discriminating screams from noise and gunshots from noise. Each classifier is trained using different features, appropriately chosen from a set of 47 audio features, which are selected according to a 2-step process. First, feature subsets of increasing size are assembled using filter selection heuristics. Then, a classifier is trained and tested with each feature subset. The obtained classification performance is used to determine the optimal feature vector dimension. This allows a noticeable speed-up w.r.t. wrapper feature selection methods. In order to validate the proposed detection algorithm, we carried out extensive experiments on a rich set of gunshots and screams mixed with ambient noise at different SNRs. Our results demonstrate that the system is able to guarantee a precision of 90% at a false rejection rate of 8%

With the increasing use of audio sensors in surveillance and monitoring applications, event detection using audio streams has emerged as an important research problem. This paper presents a hierarchical approach for audio based event detection for surveillance. The proposed approach first classifies a given audio frame into vocal and nonvocal events, and then performs further classification into normal and excited events. We model the events using a Gaussian Mixture Model and optimize the parameters for four different audio features ZCR, LPC, LPCC and LFCC. Experiments have been performed to evaluate the effectiveness of the features for detecting various normal and the excited state human activities. The results show that the proposed top-down event detection approach works significantly better than the single level approach.

Our mobility is an important daily requirement so much so that any disruption to it severely degrades our perceived quality of life. Studies in gait and human movement sciences, therefore, play a significant role in maintaining the well-being of our mobility. Current gait analysis involves numerous interdependent gait parameters that are difficult to adequately interpret due to the large volume of recorded data and lengthy assessment times in gait laboratories. A proposed solution to these problems is computational intelligence (CI), which is an emerging paradigm in biomedical engineering most notably in pathology detection and prosthesis design. The integration of CI technology in gait systems facilitates studies in disorders caused by lower limb defects, cerebral disorders, and aging effects by learning data relationships through a combination of signal processing and machine learning techniques. Learning paradigms, such as supervised learning, unsupervised learning, and fuzzy and evolutionary algorithms, provide advanced modeling capabilities for biomechanical systems that in the past have relied heavily on statistical analysis. CI offers the ability to investigate nonlinear data relationships, enhance data interpretation, design more efficient diagnostic methods, and extrapolate model functionality. These are envisioned to result in more costeffective, efficient, and easy-to-use systems, which would address global shortages in medical personnel and rising medical costs. This paper surveys current signal processing and CI methodologies followed by gait applications ranging from normal gait studies and disorder detection to artificial gait simulation. We review recent systems focusing on the existing challenges and issues involved in making them successful. We also examine new research in sensor technologies for gait that could be combined with these intelligent systems to develop more effective healthcare solutions.

This paper presents a wireless intelligent incontinence management system being developed for the University Malaya Medical Center (UMMC) that utilizes "smart" diapers to discreetly monitor and estimate wetness events, detect other relevant clinical conditions and alert staff by transmitting information via wireless technology to an intelligent central management system. We describe the hardware and software modules of the system and give results conducted on diaper experiments and the transmission range of the wireless technology.

The analysis of social media content for the extraction of geospatial information and event-related knowledge has recently received substantial attention. In this article we present an approach that leverages the complementary nature of social multimedia content by utilizing heterogeneous sources of social media feeds to assess the impact area of a natural disaster. More specifically, we introduce a novel social multimedia triangulation process that uses both Twitter and Flickr content in an integrated two-step process: Twitter content is used to identify toponym references associated with a disaster; this information is then used to provide approximate orientation for the associated Flickr imagery, allowing us to delineate the impact area as the overlap of multiple view footprints. In this approach, we practically crowdsource approximate orientations from Twitter content and use this information to orient Flickr imagery accordingly and identify the impact area through viewshed analysis and viewpoint integration. This approach enables us to avoid computationally intensive image analysis tasks associated with traditional image orientation, while allowing us to triangulate numerous images by having them pointed towards the crowdsourced toponym location. The article presents our approach and demonstrates its performance using a real-world wildfire event as a representative application case study.

A large variety of features can be extracted from raw multimedia items. Moreover, in many contexts, like in the case of multimedia uploaded by users of social media platforms, items may be linked to metadata that can be very useful for a variety of analysis tasks. Nevertheless, such features are typically heterogeneous and are difficult to combine in a unified representation that would be suitable for analysis. In this paper, we discuss the problem of clustering collections of multimedia items with the purpose of detecting social events. In order to achieve this, a novel multimodal clustering algorithm is proposed. The proposed method uses a known clustering in the currently examined domain, in order to supervise the multimodal fusion and clustering procedure. It is tested on the MediaEval social event detection challenge data and is compared to a multimodal spectral clustering approach that uses early fusion. By taking advantage of the explicit supervisory signal, it achieves superior clustering accuracy and additionally requires the specification of a much smaller number of parameters. Moreover, the proposed approach has wider scope; it is not only applicable to the task of social event detection, but to other multimodal clustering problems as well.

Driving in fog is a potentially dangerous activity that has been investigated in a number of different ways; however, most have focused on identifying the underlying perceptual changes that result in an inability to perceive speed of vehicle motion. Although the previous research has identified the perceptual changes associated with driving in fog and shows that people are highly likely to perceive their speed to be higher than it actually is, these research studies have not investigated driving behavior when drivers are allowed to maintain speed as they feel appropriate and make use of the vehicle's speedometer. In addition, much of the existing research focuses on speed perception and presents a limited view of other driving performance metrics in terms of lane keeping and event detection. The current study addresses these issues utilizing a driving simulator-based method where fog is simulated as a distance dependent contrast reduction while having participants drive at speeds they feel are appropriate. A number of different instructions and speed feedback mechanisms were tested in order to determine how drivers react when driving in varying levels of fog. Results also include lane keeping measures in order to assess whether drivers are willing to drive at speeds where their lane keeping performance is degraded due to the reduced visibility. Results indicate that, in general, drivers do not tend to slow down significantly until visibility distance is drastically reduced by fog; however, lane keeping ability is maintained throughout most of the range of visibility distances. Lane keeping ability was reduced only when fog results in visibility distances <30 m. Overall, the current study shows that drivers are willing and able to maintain vehicular control at high speed when driving in fog; however, it is important to note that drivers chose to drive at speeds where they would be incapable of stopping to avoid obstacles in the roadway even if they were to identify and react to the obstacle immediately at the border of visibility distance. This research suggests that safety benefits may be gained by convincing drivers to slow down more than they would on their own when driving in fog or enhancing a vehicle's ability to identify and react to hazards that are not visible to the driver. In order to further understand the effects of driving in fog, future naturalistic driving research should focus on identifying and mitigating risky behaviors associated with driving in foggy conditions.

This paper presents a novel approach towards automated highlight generation of broadcast sports video sequences from its extracted events and semantic concepts. A sports video is hierarchically divided into temporal partitions namely, megaslots, slots, and semantic entities, namely concepts, and events. The proposed method extracts event sequence from video and classifies each sequence into a concept by sequential association mining. The extracted concepts and events within the concepts are selected according to their degree of importance to include those in the highlights. A parameter degree of abstraction is proposed, which gives a choice to the user about how concisely the extracted concepts should be produced for a specified highlight duration. We have successfully extracted highlights from recorded video of cricket match and compared our results with the manually-generated highlights by sports television channel.

A horizontal printed Write-Once-Read-Many (WORM) resistive memory has been developed for use in wireless sensor tags targeting single-event detection in smart packaging applications. The WORM memory can be programmed using a 1.5-V printed battery. An alternative programming method is to use chemical sintering which allows the development of exposure-time triggered single event tags that can be accessed wirelessly. The new WORM memory has very low losses in the ON-state which allows direct integration into antenna structures. A sensor tag that utilizes the WORM memory functionality and the well established Electronic Article Surveillance (EAS) communication standard has been outlined. Both active and fully passive sensor tag solutions have been proposed. The role of printed electronics in smart packaging applications has been reviewed and discussed. Important enabling factors for the future development have been highlighted, such as the need for hierarchical design and test tools, better printed interconnect technologies as well as better components that allow communication with existing information and communication technology (ICT) standards. This is illustrated and underlined by the presented smart packaging concept demonstrators.

The need for efficient, sophisticated features for speech event detection is inherent in state of the art processing, enhancement and recognition systems. We explore ideas and techniques from non-linear speech modeling and analysis, like modulations and multiband filtering and propose new energy and spectral content features derived through filtering in multiple frequency bands and tracking dominant modulation energy in terms of the Teager-Kaiser Energy of separate AM-FM components. We present a detection-theoretic motivation and incorporate them in two detection schemes namely word boundary and voice activity detection. The modulation approach demonstrated noisy speech endpoint detection accuracy, reaching ∼40% error reduction on NTIMIT. In a voice activity scheme, improvement in overall misclassification error of a high hit-rate detector reached 7.5% on Aurora 2 and 9.5% on Aurora 3 databases.

A group of 32 healthy men (M) divided into three dierent age groups, i.e. M20 years [mean 21 (SD 1); n 12], M40 [mean 40 (SD 2); n 10] and M70 [mean 71 (SD 5); n 10] volunteered as subjects for examination of maximal and explosive force production of leg extensor muscles in both isometric and dynamic actions (squat jump, SJ and counter movement jump, CMJ, and standing long-jump, SLJ). The balance test was performed on a force platform in both isometric and dynamic actions. Maximal bilateral isometric force value in M70 was lower (P < 0.001) than in M40 and as much as 46% lower (P < 0.001) than that recorded in M20 (P < 0.001). The maximal rate of force development (RFD) on the force-time curve was in M70 lower (P < 0.001) than in M40 and as much as 64% lower than in M20. The heights in SJ and CMJ and the distance in SLJ in M70 were lower (P < 0.001) than in M40 and M20 (P < 0.001). In response to modi®cations of the visual surroundings the older subjects were 24%±47% (P < 0.05 and P < 0.001) slower in their response time in reaching the lit centre (TT) and remained 20%±34% (P < 0.001) less time inside the centre (TC) from the overall time of lighting than M40 and M20, respectively. In both older groups the individual values of isometric RFD correlated signi®cantly (P < 0.05) with the individual balance values of TT and TC. The present results would suggest that the capacity for explosive force production declines drastically with increasing age, even more than maximal muscle strength. Aging may also lead to impaired balance with a decrease in event detection and speed of postural adjustments. The decreased ability to develop force rapidly in older people seems to be associated with a lower capacity for neuromuscular response in controlling postural sway.

State-of-the-art statistical NLP systems for a variety of tasks learn from labeled training data that is often domain specific. However, there may be multiple domains or sources of interest on which the system must perform. For example, a spam filtering system must give high quality predictions for many users, each of whom receives emails from different sources and may make slightly different decisions about what is or is not spam. Rather than learning separate models for each domain, we explore systems that learn across multiple domains. We develop a new multidomain online learning framework based on parameter combination from multiple classifiers. Our algorithms draw from multi-task learning and domain adaptation to adapt multiple source domain classifiers to a new target domain, learn across multiple similar domains, and learn across a large number of disparate domains. We evaluate our algorithms on two popular NLP domain adaptation tasks: sentiment classification and spam filtering.

Given the complexity of healthcare and the ‘people’ nature of healthcare work and delivery, STSA (Sociotechnical Systems Analysis) research is needed to address the numerous quality of care problems observed across the world. This paper describes open STSA research areas, including workload management, physical, cognitive and macroergonomic issues of medical devices and health information technologies, STSA in transitions of care, STSA of patient-centered care, risk management and patient safety management, resilience, and feedback loops between event detection, reporting and analysis and system redesign.

E-commerce is an important information system in the network and digital age. However, the network intrusion, malicious users, virus attack and system security vulnerabilities have continued to threaten the operation of the e-commerce, making e-commerce security encounter serious test. How to improve e-commerce security has become a topic worthy of further exploration. Combining routine security test and security event detection procedures, this paper proposes the Two-Layer Secure Prevention Mechanism (TLSPM). Applying TLSPM, routine security test procedure can identify security vulnerability and defect, and develop repair operations. Security event detection procedure can timely detect security event, and assist follow repair. TLSPM can enhance the e-commerce security and effectively reduce the security risk of e-commerce critical data and asset.

Recently, wireless body area network (WBAN) plays an important role in remote cardiac patient monitoring, and mobile healthcare applications. Generally, the use of WBAN technology is restricted by size, power consumption, transmission capacity (bandwidth), and computational loads. In this paper, we therefore propose an automated cardiac event change detection for continuous remote patient monitoring devices. The proposed event change detection algorithm consists of two stages: i) ECG beat extraction; and ii) ECG beat similarity measure. In the first stage, the onset of each QRS complex is identified using the Gaussian derivative based QRS detector and the two heuristics rules. In the second stage, we employ the weighted wavelet distance (WWD) metric for finding the similarity between two ECG beats in wavelet domain. The WWD is the weighted normalized Euclidean wavelet distance between the wavelet subband coefficients vectors of the current and past ECG beats, where weights are equal to the relative wavelet subband energies of the corresponding subbands. The experimental results show that the weighted wavelet distance measure works substantially better than the conventional PRD and the wavelet based weighted PRD (WWPRD) measures under noisy environments. The proposed approach has been tested and yielded an accuracy of 99.76% on MIT-BIH Arrhythmia Database.

ly (within 1 h of randomisation) and to achieve walking by day 5 and were less likely to develop complications of immobility. The AM group was significantly (p ! 0.05) more likely to have pre-defined physiological complication events detected. All these associations remained, but were less statistically significant, after correcting for age, baseline NIH score and co-interventions. There were no significant safety concerns. Discussion: We have demonstrated the feasibility of implementing EM and AM for physiological complications in a randomised controlled trial. Larger trials are warranted to determine whether these interventions have clinical benefits.

This paper aims to address two of the key research issues in computer vision -the detection and tracking of multiple objects in the cluttered dynamic scene -that underpin the intelligence aspects of advanced visual surveillance systems aiming at automated visual events detection and behaviour analysis. We discuss two major contributions in resolving these problems within a systematic framework. Firstly, for accurate object detection, an efficient and effective scheme is proposed to remove cast shadows/highlights with error corrections based on a conditional morphological reconstruction. Secondly, for effective tracking, a temporal-template-based tracking scheme is introduced, using multiple descriptive cues (velocity, shape, colour, etc) of the 2-D object appearance together with their respective variances over time. A scaled Euclidean distance is used as the matching metric, and the template is updated using Kalman filters when a matching is found or by linear mean prediction in the case of occlusion. Extensive experiments are carried out on video sequences from various realworld scenarios. The results show very promising tracking performance.

Vast spread of sensitive loads in power systems results in increasing susceptibility to power quality problems, which makes fast detection and classification algorithms a necessity. A new approach for power quality event detection is presented in this paper. This approach utilizes linear adaptive neuron, which is called Adaline. Unlike the previously proposed detection methods, which are based on wavelet analysis that encounter mathematical calculations burden, Adaline is fast due to its simple construction, which makes it more suitable for on-line applications. #

In this study, a simple mathematical-statistical based metric called Multiple Higher Order Moments (MHOM) is introduced enabling the electrocardiogram (ECG) detection-delineation algorithm to yield acceptable results in the cases of ambulatory holter ECG including strong noise, motion artifacts, and severe arrhythmia(s). In the MHOM measure, important geometric characteristics such as maximum value to minimum value ratio, area, extent of smoothness or being impulsive and distribution skewness degree (asymmetry), occult. In the proposed method, first three leads of high resolution 24-h holter data are extracted and preprocessed using Discrete Wavelet Transform (DWT). Next, a sample to sample sliding window is applied to preprocessed sequence and in each slid, mean value, variance, skewness, and kurtosis of the excerpted segment are superimposed called MHOM. The MHOM metric is then used as decision statistic to detect and delineate ECG events. To show advantages of the presented method, it is applied to MIT-BIH Arrhythmia Database, QT Database, and T-Wave Alternans Database and as a result, the average values of sensitivity and positive predictivity Se = 99.95% and P+ = 99.94% are obtained for the detection of QRS complexes, with the average maximum delineation error of 6.1, 4.1, and 6.5 ms for P-wave, QRS complex, and T-wave, respectively showing marginal improvement of detectiondelineation performance. In the next step, the proposed method is applied to DAY hospital high resolution holter data (more than 1,500,000 beats including Bundle Branch Blocks-BBB, Premature Ventricular Complex-PVC, and Premature Atrial Complex-PAC) and average values of Se = 99.97% and P+ = 99.95% are obtained for QRS detection. In summary, marginal performance improvement of ECG events detection-delineation process, reliable robustness against strong noise, artifacts, and probable severe arrhythmia(s) of high resolution holter data can be mentioned as important merits and capabilities of the proposed algorithm.

Twitter has received much attention recently. An important characteristic of Twitter is its real-time nature. We investigate the real-time interaction of events such as earthquakes in Twitter and propose an algorithm to monitor tweets and to detect a target event. To detect a target event, we devise a classifier of tweets based on features such as the keywords in a tweet, the number of words, and their context. Subsequently, we produce a probabilistic spatiotemporal model for the target event that can find the center of the event location. We regard each Twitter user as a sensor and apply particle filtering, which are widely used for location estimation. The particle filter works better than other comparable methods for estimating the locations of target events. As an application, we develop an earthquake reporting system for use in Japan. Because of the numerous earthquakes and the large number of Twitter users throughout the country, we can detect an earthquake with high probability (93 percent of earthquakes of Japan Meteorological Agency (JMA) seismic intensity scale 3 or more are detected) merely by monitoring tweets. Our system detects earthquakes promptly and notification is delivered much faster than JMA broadcast announcements.

The conditions of networked ensemble performance were sim-ulated in an experiment. Pairs of musicians were placed apart in isolated rooms and given a simple rhythm to clap together. A microphone was placed as close as possible to each per-former&#x27;s hands. ...

Given the complexity of health care and the 'people' nature of healthcare work and delivery, STSA (Sociotechnical Systems Analysis) research is needed to address the numerous quality of care problems observed across the world. This paper describes open STSA research areas, including workload management, physical, cognitive and macroergonomic issues of medical devices and health information technologies, STSA in transitions of care, STSA of patientcentered care, risk management and patient safety management, resilience, and feedback loops between event detection, reporting and analysis and system redesign.

Event detection is used to classify recorded gaze points into periods of fixation, saccade, smooth pursuit, blink, and noise. Although there is an overall consensus that current algorithms for event detection have serious flaws and that a de facto standard for event detection does not exist, surprisingly little work has been done to remedy this problem. We suggest a new velocity-based algorithm that takes several of the previously known limitations into account. Most important, the new algorithm identifies so-called glissades, a wobbling movement at the end of many saccades, as a separate class of eye movements. Part of the solution involves designing an adaptive velocity threshold that makes the event detection less sensitive to variations in noise level and the algorithm settings-free for the user. We demonstrate the performance of the new algorithm on eye movements recorded during reading and scene perception and compare it with two of the most commonly used algorithms today. Results show that, unlike the currently used algorithms, fixations, saccades, and glissades are robustly identified by the new algorithm. Using this algorithm, we found that glissades occur in about half of the saccades, during both reading and scene perception, and that they have an average duration close to 24 msec. Due to the high prevalence and long durations of glissades, we argue that researchers must actively choose whether to assign the glissades to saccades or fixations; the choice affects dependent variables such as fixation and saccade duration significantly. Current algorithms do not offer this choice, and their assignments of each glissade are largely arbitrary.

Automatic detection of a falling person in video sequences has interesting applications in video -surveillance and is an important part of future pervasive home monitoring systems. In this paper, we propose a multiview approach to achieve this goal, where motion is modeled using a layered hidden Markov model (LHMM). The posture classification is performed by a fusion unit, merging the decision provided by the independently processing cameras in a fuzzy logic context. In each view, the fall detection is optimized in a given plane by performing a metric image rectification, making it possible to extract simple and robust features, and being convenient for real-time purpose. A theoretical analysis of the chosen descriptor enables us to define the optimal camera placement for detecting people falling in unspecified situations, and we prove that two cameras are sufficient in practice. Regarding event detection, the LHMM offers a principle way for solving the inference problem. Moreover, the hierarchical architecture decouples the motion analysis into different temporal granularity levels, making the algorithm able to detect very sudden changes, and robust to low-level steps errors.

We study the electroencephalogram ͑EEG͒ of 30 closed-eye awake subjects with a technique of analysis recently proposed to detect punctual events signaling rapid transitions between different metastable states. After single-EEG-channel event detection, we study global properties of events simultaneously occurring among two or more electrodes termed coincidences. We convert the coincidences into a diffusion process with three distinct rules that can yield the same only in the case where the coincidences are driven by a renewal process. We establish that the time interval between two consecutive renewal events driving the coincidences has a waiting-time distribution with inverse power-law index Ϸ 2 corresponding to ideal 1 / f noise. We argue that this discovery, shared by all subjects of our study, supports the conviction that 1 / f noise is an optimal communication channel for complex networks as in art or language and may therefore be the channel through which the brain influences complex processes and is influenced by them.

Event-based prospective memory (PM) requires remembering the delayed execution of an intended action in response to a pre-specified PM cue while being actively engaged in an ongoing task in which the cue is embedded. To date, experimental paradigms vary as to whether or not they require participants immediately to stop working on the ongoing task whenever they encounter a PM event (cue) and directly switch to the prospective action (task-switch approach). Alternatively, several other paradigms used in the literature encourage participants to continue working on the ongoing task item after the cue, and only then, perform the prospective action (dual-task approach). The present study explores the possible behavioural and electrophysiological effects that both approaches may have on PM performance. Seventeen young adults performed both versions of a standard PM task in a counterbalanced order during which behavioural data and electroencephalogram (EEG) were recorded. Behavioural data showed a decrement in PM performance in the task-switch compared to the dual-task condition. In addition, EEG data revealed differences between the dual-task and task-switch approach in event-related potential (ERP) components associated with response inhibition and with post-retrieval monitoring (i.e. late positive complex). No differences between the two tasks were found with regard to the PM event detection processes (i.e. N300) and the retrieval of the intended action from long-term memory. In sum, findings demonstrate that it does make a difference which task approach is applied and suggest that dual-task and task-switch paradigms may result in different processing and neurophysiological dynamics particularly concerning attentional resources and cognitive control.

With digital still cameras, users can easily collect thousands of photos. Our goal is to make organizing and browsing photos simple and quick, while retaining scalability to large collections. To that end, we created a photo management application concentrating on areas that improve the overall experience without neglecting the mundane components of such an application. Our application automatically divides photos into meaningful events such as birthdays or trips. Several user interaction mechanisms enhance the user experience when organizing photos. Our application combines a light table for showing thumbnails of the entire photo collection with a tree view that supports navigating, sorting, and filtering photos by categories such as dates, events, people, and locations. A calendar view visualizes photos over time and allows for the quick assignment of dates to scanned photos. We fine-tuned our application by using it with large personal photo collections provided by several users.

The analysis of social media content for the extraction of geospatial information and event-related knowledge has recently received substantial attention. In this article we present an approach that leverages the complementary nature of social multimedia content by utilizing heterogeneous sources of social media feeds to assess the impact area of a natural disaster. More specifically, we introduce a novel social multimedia triangulation process that uses both Twitter and Flickr content in an integrated two-step process: Twitter content is used to identify toponym references associated with a disaster; this information is then used to provide approximate orientation for the associated Flickr imagery, allowing us to delineate the impact area as the overlap of multiple view footprints. In this approach, we practically crowdsource approximate orientations from Twitter content and use this information to orient Flickr imagery accordingly and identify the impact area through viewshed analysis and viewpoint integration. This approach enables us to avoid computationally intensive image analysis tasks associated with traditional image orientation, while allowing us to triangulate numerous images by having them pointed towards the crowdsourced toponym location. The article presents our approach and demonstrates its performance using a real-world wildfire event as a representative application case study.

This is a project report/summary of TrioVecEvent paper for local event detection from twitter feed.

This paper was orignally published by Chao Zhang, Liyuan Liu, Dongming Lei, Qan Yuan, Honglei Zhuang, Tim Hanratty, and Jiawei Han at 23rd KDD (Knowledge, Discovery and Data Mining) international conference in August 2017.

One of the critical success factors of event-driven systems is the capability of detecting complex events from simple and ordinary event notifications. Complex events which trigger or terminate actionable situations can be inferred from large event clouds or event streams based on their event instance sequence, their syntax and semantics. Using semantics of event algebra patterns defined on top of event instance sequences for event detection is one of the promising approaches for detection of complex events. The developments and successes in building standards and tools for semantic technologies such as declarative rules and ontologies are opening novel research and application areas in event processing. One of these promising application areas is semantic event processing. In this paper we contribute with a conceptual approach which supports the implementation of the vision of semantic event-driven systems; using Semantic Web technologies, benefiting from complex event processing, and ensuring quality through trust and reputation management. All of these novel technologies leads to more intelligent decision supporting systems. 1

Event-based prospective memory (PM) requires remembering the delayed execution of an intended action in response to a pre-specified PM cue while being actively engaged in an ongoing task in which the cue is embedded. To date, experimental paradigms vary as to whether or not they require participants immediately to stop working on the ongoing task whenever they encounter a PM event (cue) and directly switch to the prospective action (task-switch approach). Alternatively, several other paradigms used in the literature encourage participants to continue working on the ongoing task item after the cue, and only then, perform the prospective action (dual-task approach). The present study explores the possible behavioural and electrophysiological effects that both approaches may have on PM performance. Seventeen young adults performed both versions of a standard PM task in a counterbalanced order during which behavioural data and electroencephalogram (EEG) were recorded. Behavioural data showed a decrement in PM performance in the task-switch compared to the dual-task condition. In addition, EEG data revealed differences between the dual-task and task-switch approach in event-related potential (ERP) components associated with response inhibition and with post-retrieval monitoring (i.e. late positive complex). No differences between the two tasks were found with regard to the PM event detection processes (i.e. N300) and the retrieval of the intended action from long-term memory. In sum, findings demonstrate that it does make a difference which task approach is applied and suggest that dual-task and task-switch paradigms may result in different processing and neurophysiological dynamics particularly concerning attentional resources and cognitive control.

We present an attentional selection system for processing video streams from remotely operated underwater vehicles (ROVs). The system identifies potentially interesting visual events spanning multiple frames based on low-level spatial properties of salient tokens, which are associated with those events and tracked over time. If video frames contain interesting events, they are labeled "interesting", otherwise they are labeled "boring". By marking the "interesting" events and omitting "boring" frames in the output stream, we augment the productivity of human video annotators, or, alternatively, provide input for a subsequent object classification algorithm.

Rapid event detection faces an emergent need to process large videos collections; whether surveillance videos or unconstrained web videos, the ability to automatically recognize high-level, complex events is a challenging task. Motivated by pre-existing methods being complex, computationally demanding, and often non-replicable, we designed a simple system that is quick, effective and carries minimal overhead in terms of memory and storage. Our system is clearly described, modular in nature, replicable on any Desktop, and demonstrated with extensive experiments, backed by insightful analysis on different Convolutional Neural Networks (CNNs), as stand-alone and fused with others. With a large corpus of unconstrained, real-world video data, we examine the usefulness of different CNN models as features extractors for mod-eling high-level events, i.e., pre-trained CNNs that differ in architectures, training data, and number of outputs. For each CNN, we use 1-fps from all training exemplar to train one-vs-rest SVMs for each event. To represent videos, frame-level features were fused using a variety of techniques. The best being to max-pool between predetermined shot boundaries, then average-pool to form the final video-level descriptor. Through extensive analysis, several insights were found on using pre-trained CNNs as off-the-shelf feature extractors for the task of event detection. Fusing SVMs of different CNNs revealed some interesting facts, finding some combinations to be complimentary. It was concluded that no single CNN works best for all events, as some events are more object-driven while others are more scene-based. Our top performance resulted from learning event-dependent weights for different CNNs.

The Elat fault (a segment of the Dead Sea Transform) runs along the southern Arava valley (part of the Dead Sea Rift, Israel) forming a complex fault zone that displays a time-dependent seismic behaviour. Paleoseismic evidence shows that this fault zone has generated at least 15 earthquakes of magnitude larger than M 6 during the late Pleistocene and the Holocene. However, at present the Elat fault is one of the quietest segments of the Dead Sea Transform, lacking even microsesimicity. The last event detected in the southern Arava valley occurred in the Avrona playa and was strong enough to have deformed the playa and to change it from a closed basin with internal drainage into an open basin draining to the south.