Amir H. Meghdadi

Interactive Exploration of Surveillance Video through Action Shot Summarization and Trajectory Visualization

We propose a novel video visual analytics system for interactive exploration of surveillance video data. Our approach consists of providing analysts with various views of information related to moving objects in a video. To do this we first extract each objects movement path. We visualize each movement by (a) creating a single action shot image (a still image that coalesces multiple frames), (b) plotting its trajectory in a space-time cube and (c) displaying an overall timeline view of all the movements. The action shots provide a still view of the moving object while the path view presents movement properties such as speed and location. We also provide tools for spatial and temporal filtering based on regions of interest. This allows analysts to filter out large amounts of movement activities while the action shot representation summarizes the content of each movement. We incorporated this multi-part visual representation of moving objects in sViSIT, a tool to facilitate browsing through the video content by interactive querying and retrieval of data. Based on our interaction with security personnel who routinely interact with surveillance video data, we identified some of the most common tasks performed. This resulted in designing a user study to measure time-to-completion of the various tasks. These generally required searching for specific events of interest (targets) in videos. Fourteen different tasks were designed and a total of 120 min of surveillance video were recorded (indoor and outdoor locations recording movements of people and vehicles). The time-to-completion of these tasks were compared against a manual fast forward video browsing guided with movement detection. We demonstrate how our system can facilitate lengthy video exploration and significantly reduce browsing time to find events of interest. Reports from expert users identify positive aspects of our approach which we summarize in our recommendations for future video visual analytics systems.

Tolerance Classes in Measuring Image Resemblance

The problem considered in this paper is how to measure resemblance between images. One approach to the solution to this problem is to find parts of images that resemble each other with a tolerable level of error. This leads to a consideration of tolerance relations that define coverings of images and measurement of the degree of overlap between tolerances classes in pairs of images. This approach is based on a tolerance class form of near sets that model human perception in a physical continuum. This is a humanistic perception-based near set approach, where tolerances become part of the solution to the image correspondence problem. Near sets are a generalization of rough sets introduced by Zdzislaw Pawlak during the early 1980s. The basic idea in devising near set-based measures of resemblance of images that emulate human perception is to allow overlapping classes in image coverings defined with respect to a tolerance *** . The contribution of this article is the introduction of two new tolerance class-based image resemblance measures and a comparison of the new measures with the original Henry-Peters image nearness measure.

Nature-inspired framework for measuring visual image resemblance: A near rough set approach

The problem considered in this paper is how to determine the degree of nearness between complex visual objects. The proposed solution to this problem stems from a natural computing approach to solving the visual acuity problem in terms of a granular representation of visual information that is quantifiable as well as understandable for humans. This is accomplished via a near rough set framework in the approximation of a pair of disjoint sets and measurement of distances between sets using various fuzzy pseudometrics. Pseudometrics, in general, and fuzzy pseudometrics, in particular, are useful in measuring the distance between pairs of objects such as sets. Such distances are indicators of the nearness of (resemblance between) visual objects. These observations lead to a number of practical applications such as object recognition and object retrieval in digital image analysis. One such application is reported in this article. The contribution of this article is threefold: introduction of a nature-inspired framework for measurement of visual object resemblance, four different incarnations of the standard fuzzy metric and application of fuzzy metrics in content-based image retrieval experiments.

Measuring Resemblances Between Swarm Behaviours: A Perceptual Tolerance Near Set Approach

The problem considered in this article is how to detect and measure resemblances between swarm behaviours. The solution to this problem stems from an extension of recent work on tolerance near sets and image correspondence. Instead of considering feature extraction from subimages in digital images, we compare swarm behaviours by considering feature extraction from subsets of tuples of feature-values representing the behaviour of observed swarms of organisms. Thanks to recent work on the foundations of near sets, it is possible to formulate a rigorous approach to measuring the extent that swarm behaviours resemble each other. Fundamental to this approach is what is known as a recent description-based set intersection, a set containing objects with matching or almost the same descriptions extracted from objects contained in pairs of disjoint sets. Implicit in this work is a new approach to comparing information tables representing N. Tinbergens ethology (study of animal behaviour) and direct result of recent work on what is known as rough ethology. Included in this article is a comparison of recent nearness measures that includes a new form of F. Hausdorffs distance measure. The contribution of this article is a tolerance near set approach to measuring the degree of resemblance between swarm behaviours.

Detecting Determinism in EEG Signals using Principal Component Analysis and Surrogate Data Testing

A novel method is proposed here to determine whether a time series is deterministic even in the presence of noise. The method is the extension of an existing method based on smoothness analysis of the signal in state space with surrogate data testing. While classical measures fail to detect determinism when the time series is corrupted by noise, the proposed method can clearly distinguish between pure stochastic and originally deterministic but noisy time series. A set of measures is defined here named partial smoothness indexes corresponding to principal components of the time series in state space. It is shown that when the time series is not pure stochastic, at least one of the indexes reflects determinism. The method is first successfully tested through simulation on a chaotic Lorenz time series contaminated with noise and then applied on EEG signals. Testing results on both our experimental recorded EEG signals and a benchmark EEG database verifies this hypothesis that EEG signals are deterministic in nature while contain some stochastic components as well

Characterization of healthy and epileptic brain EEG signals by monofractal and multifractal analysis

This paper presents a method based on fractal dimensions to characterize electroencephalogram (EEG) signals, and differentiate between healthy and epileptic EEG data sets. The estimated correlation fractal dimension is considerably lower for intracranial invasive EEG recordings as compared to non-invasive scalp recordings. The epileptic EEG is also shown to have lower correlation dimension than healthy EEG. Multifractal analysis of EEG signal using the Renyi fractal dimension spectrum also demonstrates lower absolute values and variability in the spectrum for seizure activity compared to normal brain activity. Finally, a moving window scheme is utilized to analyze EEG signal prior to epileptic seizures in search for a pattern to predict an impending seizure. The results of the later study are not conclusive at this point yet.

Perceptual tolerance neighborhood‐based similarity in content‐based image retrieval and classification

Purpose – The purpose of this paper is to demonstrate the effectiveness and advantages of using perceptual tolerance neighbourhoods in tolerance space‐based image similarity measures and its application in content‐based image classification and retrieval.Design/methodology/approach – The proposed method in this paper is based on a set‐theoretic approach, where an image is viewed as a set of local visual elements. The method also includes a tolerance relation that detects the similarity between pairs of elements, if the difference between corresponding feature vectors is less than a threshold 2 (0,1).Findings – It is shown that tolerance space‐based methods can be successfully used in a complete content‐based image retrieval (CBIR) system. Also, it is shown that perceptual tolerance neighbourhoods can replace tolerance classes in CBIR, resulting in more accuracy and less computations.Originality/value – The main contribution of this paper is the introduction of perceptual tolerance neighbourhoods instead o...

A Method for Detecting Nonlinear Determinism in Normal and Epileptic Brain EEG Signals

A robust method of detecting determinism for short time series is proposed and applied to both healthy and epileptic EEG signals. The method provides a robust measure of determinism through characterizing the trajectories of the signal components which are obtained through singular value decomposition. Robustness of the method is shown by calculating proposed index of determinism at different levels of white and colored noise added to a simulated chaotic signal. The method is shown to be able to detect determinism at considerably high levels of additive noise. The method is then applied to both intracranial and scalp EEG recordings collected in different data sets for healthy and epileptic brain signals. The results show that for all of the studied EEG data sets there is enough evidence of determinism. The determinism is more significant for intracranial EEG recordings particularly during seizure activity.