Efficient temporal mining of micro-blog texts and its application to event discovery

In this paper we present a novel method for clustering words in micro-blogs, based on the similarity of the related temporal series. Our technique, named SAX*, uses the Symbolic Aggregate ApproXimation algorithm to discretize the temporal series of terms into a small set of levels, leading to a string for each. We then define a subset of “interesting” strings, i.e. those representing patterns of collective attention. Sliding temporal windows are used to detect co-occurring clusters of tokens with the same or similar string. To assess the performance of the method we first tune the model parameters on a 2-month 1 % Twitter stream, during which a number of world-wide events of differing type and duration (sports, politics, disasters, health, and celebrities) occurred. Then, we evaluate the quality of all discovered events in a 1-year stream, “googling” with the most frequent cluster n-grams and manually assessing how many clusters correspond to published news in the same temporal slot. Finally, we perform a complexity evaluation and we compare SAX* with three alternative methods for event discovery. Our evaluation shows that SAX* is at least one order of magnitude less complex than other temporal and non-temporal approaches to micro-blog clustering.     

A multivariate Bayesian scan statistic for early event detection and characterization

We present the multivariate Bayesian scan statistic (MBSS), a general framework for event detection and characterization in multivariate spatial time series data. MBSS integrates prior information and observations from multiple data streams in a principled Bayesian framework, computing the posterior probability of each type of event in each space-time region. MBSS learns a multivariate Gamma-Poisson model from historical data, and models the effects of each event type on each stream using expert knowledge or labeled training examples. We evaluate MBSS on various disease surveillance tasks, detecting and characterizing outbreaks injected into three streams of Pennsylvania medication sales data. We demonstrate that MBSS can be used both as a “general” event detector, with high detection power across a variety of event types, and a “specific” detector that incorporates prior knowledge of an event’s effects to achieve much higher detection power. MBSS has many other advantages over previous event detection approaches, including faster computation and easy interpretation and visualization of results, and allows faster and more accurate event detection by integrating information from the multiple streams. Most importantly, MBSS can model and differentiate between multiple event types, thus distinguishing between events requiring urgent responses and other, less relevant patterns in the data.     

A fast connected components labeling algorithm and its application to real-time pupil detection

We describe a fast connected components labeling algorithm using a region coloring approach. It computes region attributes such as size, moments, and bounding boxes in a single pass through the image. Working in the context of real-time pupil detection for an eye tracking system, we compare the time performance of our algorithm with a contour tracing-based labeling approach and a region coloring method developed for a hardware eye detection system. We find that region attribute extraction performance exceeds that of these comparison methods. Further, labeling each pixel, which requires a second pass through the image, has comparable performance.     

A rule dynamics approach to event detection in Twitter with its application to sports and politics

The increasing popularity of Twitter as social network tool for opinion expression as well as information retrieval has resulted in the need to derive computational means to detect and track relevant topics/events in the network. The application of topic detection and tracking methods to tweets enable users to extract newsworthy content from the vast and somehow chaotic Twitter stream. In this paper, we apply our technique named Transaction-based Rule Change Mining to extract newsworthy hashtag keywords present in tweets from two different domains namely; sports (The English FA Cup 2012) and politics (US Presidential Elections 2012 and Super Tuesday 2012). Noting the peculiar nature of event dynamics in these two domains, we apply different time-windows and update rates to each of the datasets in order to study their impact on performance. The performance effectiveness results reveal that our approach is able to accurately detect and track newsworthy content. In addition, the results show that the adaptation of the time-window exhibits better performance especially on the sports dataset, which can be attributed to the usually shorter duration of football events.     

A general framework for time granularity and its application to temporal reasoning

This paper presents a general framework to define time granularity systems. We identify the main dimensions along which different systems can be characterized, and investigate the formal relationships among granularities in these systems. The paper also introduces the notion of a network of temporal constraints with (multiple) granularities emphasizing the semantic and computational differences from constraint networks with a single granularity. Consistency of networks with multiple granularities is shown to beNP‐hard in general and approximate solutions for this problem and for the minimal network problem are proposed. This revised version was published online in June 2006 with corrections to the Cover Date.     

A multiresolution wavelet networks architecture and its application to pattern recognition

This paper aims at addressing a challenging research in both fields of the wavelet neural network theory and the pattern recognition. A novel architecture of the wavelet network based on the multiresolution analysis (MRWN) and a novel learning algorithm founded on the Fast Wavelet Transform (FWTLA) are proposed. FWTLA has numerous positive sides compared to the already existing algorithms. By exploiting this algorithm to learn the MRWN, we suggest a pattern recognition system (FWNPR). We show firstly its classification efficiency on many known benchmarks and then in many applications in the field of the pattern recognition. Extensive empirical experiments are performed to compare the proposed methods with other approaches.     

A cusum change-point detection algorithm for non-stationary sequences with application to data network surveillance

We adapt the classic cusum change-point detection algorithm to handle non-stationary sequences that are typical with network surveillance applications. The proposed algorithm uses a defined timeslot structure to take into account time varying distributions, and uses historical samples of observations within each timeslot to facilitate a nonparametric methodology. Our proposed solution includes an on-line screening feature that fully automates the implementation of the algorithm and eliminates the need for manual oversight up until the point where root cause analysis begins.     

A unified method for evaluating real-time computer controllers and its application

A computer-controlled system is a synergistic coupling of the controlled process and the controller computer. We have defined new performance measures for real-time controller computers based on this coupling. We present a systematic study of a typical critical controlled process in the context of new performance measures that express the performance of both controlled processes and controller computers (taken as a unit) on the basis of a single variable: controller response time. Controller response time is a function of current system state, system failure rate, electrical and/or magnetic interference, etc., and is therefore a random variable. Control overhead is expressed as monotonically nondecreasing function of the response time and the system suffers catastrophic failure, or dynamic failure, if the response time for a control task exceeds the corresponding system hard deadline, if any. The controlled-process chosen for study is an aircraft in the final stages of descent, just prior to landing. Control constraints are particularly severe during this period, and great care must be taken in the design of controllers that handle this process. First, the performance measures for the controller are presented. Second, control algorithms for solving the landing problem are discussed, and finally the impact of our performance measures on the problem is analyzed, showing that the performance measures and the associated estimation method have potential use for designing and/or evaluating real-time controllers and controlled process. In common with all other control techniques, the computational complexity involved in obtaining these measures is susceptible to the curse of dimensionality.     

Boosted string representation and its application to video surveillance

This paper presents a new behavior analysis system for analyzing human movements via a boosted string representation. First of all, we propose a triangulation-based method to transform each action sequence into a set of symbols. Then, an action sequence can be interpreted and analyzed using this string representation. To analyze action sequences with this string representation, three practical problems should be tackled. Usually, an action sequence has different temporal scaling changes, different initial states, and symbol converting errors. Traditional methods (like hidden Markov models and finite state machines) have limited abilities to deal with the above problems since many unknown states should be constructed and initialized. To tackle the problems, a novel string hypothesis generator is then proposed for generating a bank of string features from which different invariant features can be learned for classifying behaviors more accurately. To learn the invariant features, the Adaboost algorithm is used and modified to train a strong classifier from the set of string hypotheses so that multiple human action events can be well classified. In addition, a forward classification scheme is proposed to classify all input action sequences more accurately even though they have various scaling changes and coding errors. Experimental results prove that the proposed method is a robust, accurate, and powerful tool for human movement analysis.     

A high speed roller dung beetles clustering algorithm and its architecture for real-time image segmentation

Applied Intelligence - Several practical applications like disaster detection, remote surveillance, object recognition using remote sensing satellite images, object monitoring and tracking using...     

Bayesian CAR models for syndromic surveillance on multiple data streams: Theory and practice

Syndromic surveillance has, so far, considered only simple models for Bayesian inference. This paper details the methodology for a serious, scalable solution to the problem of combining symptom data from a network of US hospitals for early detection of disease outbreaks. The approach requires high-end Bayesian modeling and significant computation, but the strategy described in this paper appears to be feasible and offers attractive advantages over the methods that are currently used in this area. The method is illustrated by application to ten quarters worth of data on opioid drug abuse surveillance from 636 reporting centers, and then compared to two other syndromic surveillance methods using simulation to create known signal in the drug abuse database.     

A Bayesian approach to the Hough transform for line detection

This paper explains how to associate a rigorous probability value to the main straight line features extracted from a digital image. A Bayesian approach to the Hough Transform (HT) is considered. Under general conditions, it is shown that a probability measure is associated to each line extracted from the HT. The proposed method increments the HT accumulator in a probabilistic way: first calculating the uncertainty of each edge point in the image and then using a Bayesian probabilistic scheme for fusing the probability of each edge point and calculating the line feature probability.     

A temporal logic approach to discrete event control for the safety canonical class

This paper presents a temporal logic formulation of discrete event control which forms a new theoretical basis for control analysis and synthesis of a class of discrete event systems (DES). Based on the formulation, a basic supervisory control theory is developed for a control objective specified by an invariance formula belonging to the safety canonical class of Manna and Pneuli. Using the safety canonical class as a basis, the refinement and generalization of the existing basic predicate framework are demonstrated. A simple example illustrates the formal axiomatic means to perform control-theoretic analysis and synthesis under the new formulation.     

A characterization of parity space and its application to robustfault detection

Deals with problems related to the parity relation-based residual generation. A characterization of parity vectors and a relationship between the order of parity relations and the dimension of the parity space are derived. The achieved results are used to determine the degree of freedom for designing parity relation-based residual generators and to study the robustness problem     

A real-time object detecting and tracking system for outdoor night surveillance

Autonomous video surveillance and monitoring has a rich history. Many deployed systems are able to reliably track human motion in indoor and controlled outdoor environments. However, object detection and tracking at night remain very important problems for visual surveillance. The objects are often distant, small and their signatures have low contrast against the background. Traditional methods based on the analysis of the difference between successive frames and a background frame will do not work. In this paper, a novel real time object detection algorithm is proposed for night-time visual surveillance. The algorithm is based on contrast analysis. In the first stage, the contrast in local change over time is used to detect potential moving objects. Then motion prediction and spatial nearest neighbor data association are used to suppress false alarms. Experiments on real scenes show that the algorithm is effective for night-time object detection and tracking.     

A real-time framework for eye detection and tracking

Epidemiological studies indicate that automobile drivers from varying demographics are confronted by difficult driving contexts such as negotiating intersections, yielding, merging and overtaking. We aim to detect and track the face and eyes of the driver during several driving scenarios, allowing for further understanding of a driver’s visual search pattern behavior. Traditionally, detection and tracking of objects in visual media has been performed using specific techniques. These techniques vary in terms of their robustness and computational cost. This research proposes a real-time framework that is built upon a foundation synonymous to boosting, which we extend from learners to trackers and demonstrate that the idea of an integrated framework employing multiple trackers is advantageous in forming a globally strong tracking methodology. In order to model the effectiveness of trackers, a confidence parameter is introduced to help minimize the errors produced by incorrect matches and allow more effective trackers with a higher confidence value to correct the perceived position of the target.     

Selective temporal filtering and its application to hand gesture recognition

In temporal data analysis, noisy data is inevitable in both testing and training. This noise can seriously influence the performance of the temporal data analysis. To address this problem, we propose a novel method, termed Selective Temporal Filtering that builds a noise-free model for classification during training and identifies key-feature vectors that are noise-filtered data from the input sequence during testing. The use of these key-feature vectors makes the classifier robust to noise within the input space. The proposed method is validated on a synthetic-dataset and a database of American Sign Language. Using key-feature vectors results in robust performance with respect to the noise content. Futhermore, we are able to show that the proposed method not only outperforms Conditional Random Fields and Hidden Markov Models in noisy environments, but also in a well-controlled environment where we assume no significant noise vectors exist.     

Exploiting visual quasi-periodicity for real-time chewing event detection using active appearance models and support vector machines

Steady increases in healthcare costs and obesity have inspired recent studies into cost-effective, assistive systems capable of monitoring dietary habits. Few researchers, though, have investigated the use of video as a means of monitoring dietary activities. Video possesses several inherent qualities, such as passive acquisition, that merits its analysis as an input modality for such an application. To this end, we propose a method to automatically detect chewing events in surveillance video of a subject. Firstly, an Active Appearance Model (AAM) is used to track a subject’s face across the video sequence. It is observed that the variations in the AAM parameters across chewing events demonstrate a distinct periodicity. We utilize this property to discriminate between chewing and non-chewing facial actions such as talking. A feature representation is constructed by applying spectral analysis to a temporal window of model parameter values. The estimated power spectra subsequently undergo non-linear dimensionality reduction. The low-dimensional embedding of the power spectra are employed to train a binary Support Vector Machine classifier to detect chewing events. To emulate the gradual onset and offset of chewing, smoothness is imposed over the class predictions of neighboring video frames in order to deter abrupt changes in the class labels. Experiments are conducted on a dataset consisting of 37 subjects performing each of five actions, namely, open- and closed-mouth chewing, clutter faces, talking, and still face. Experimental results yielded a cross-validated percentage agreement of 93.0%, indicating that the proposed system provides an efficient approach to automated chewing detection.