Signal detection using third-order moments

Delaney, P.A.

doi:10.1007/bf01183743

Cited by 11 publications

(10 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It should be emphasized that, even if the measurement noise is not strictly Gaussian, kurtosis is a robust detector of outliers, as shown analytically in [10], [26], [19], [30]. Thus, it can be used to reliably detect at which pixels motion is present by detecting outliers in the inter-frame illumination variations, as verified by our experiments, in Sec.…”

Section: Kurtosis-based Activity Area Localizationmentioning

confidence: 69%

“…In [10] it is rigorously proven that the kurtosis is a robust detector of outliers in Gaussian noise, but that it can also detect them when they are embedded in non-Gaussian noise. In order to demonstrate this, we consider the case of nonGaussian, zero-mean (without loss of generality, since the mean can be subtracted from our data set) additive noise v, added to a Gaussian random variable y:…”

Section: Kurtosis-based Activity Area Localizationmentioning

confidence: 99%

See 1 more Smart Citation

Human activity localization via sequential change detection

Briassouli

Kompatsiaris

2008

Proceedings of the 1st ACM International Conference on Multimedia Information Retrieval

View full text Add to dashboard Cite

Today's rapid developments in digital media processing capabilities, and network speeds, make the dissemination of multimedia data extremely rapid and reliable, and have attracted significant research attention to video analysis, event detection, tracking and surveillance. In this work, a novel, generally applicable approach to the detection of human activity in video is presented. The areas of activity in the video are first detected via the accumulation and statistical processing of the motion vectors in all frames. The times (frames) at which events begin and end are defined as moments at which the statistical distribution of the motion vectors changes, for each pixel. These time instants are estimated in a novel manner, by applying sequential likelihood ratio testing on the motion vectors of the pixels that have been found to be active.The proposed system provides a theoretically sound solution for the detection of temporal changes in the human (or other) activity in video, without resorting to use of prior knowledge, heuristics, or ad-hoc thresholds. Sequential detection techniques allow us to find the frames where events begin and end, but also allows to pre-define the desired probabilities of false alarm and miss for the system. This is entirely novel for the temporal localization of activities and events in the video processing literature. Finally, sequential change detection methods require the smallest number of samples to detect a change, so they ensure the fastest detection of events. Experiments are performed with real sequences, involving human activities, for varying probabilities of false alarm and miss. Comparison with ground truth results shows that, indeed, the proposed method leads to meaningful localization of events both in time and in space.

show abstract

Section: Kurtosis-based Activity Area Localizationmentioning

confidence: 69%

Section: Kurtosis-based Activity Area Localizationmentioning

confidence: 99%

Human activity localization via sequential change detection

Briassouli

Kompatsiaris

2008

Proceedings of the 1st ACM International Conference on Multimedia Information Retrieval

View full text Add to dashboard Cite

show abstract

“…The kurtosis is chosen to detect areas of motion in video sequences, as it has been shown to be a more robust detector (Guo et al, 1999) than simple differencing, which can suppress noise from Gaussian, and even nonGaussian data, as shown analytically by Delaney (1994). Thus, it can be used in our application even for the cases where the noise in the optical flow estimates deviates from the Gaussian model assumption.…”

Section: Motion Analysis: Activity Area Extraction From Optical Flowmentioning

confidence: 97%

Color aided motion‐segmentation and object tracking for video sequences semantic analysis

Briassouli

Mezaris

Kompatsiaris

2007

Int J Imaging Syst Tech

View full text Add to dashboard Cite

ABSTRACT:The high rates at which digital multimedia is being generated and used makes it necessary to develop systems that can process it in an efficient manner. This can be achieved by extracting semantics from processing the video's low-level information. We present a novel algorithm which fuses color and motion information, in order to extract semantics from the video sequence. The motion estimates are processed statistically to give areas of activity in the video. Color segmentation is applied to these areas, and also to their complementary regions in each frame, in order to achieve the moving object segmentation. The extracted color layers in the activity and background areas are compared using the earth mover's distance (EMD), and a novel method, which we introduce, and which is based on a likelihood ratio test (LRT). The segmentation results of our LRTbased approach are shown to be more robust than the EMD results, and both methods are shown to be more accurate than the existing combined color-motion approaches. Furthermore, the LRT method allows the retrieval of additional semantics, namely of ''maps'' that indicate with what likelihood a pixel belongs to a moving object. The areas of activity can be used to retrieve semantics for the kind of activity taking place. The color-aided segmentation of the moving entities provides a full description of their appearance, so it can be used, for example, to classify the video based on the objects in it. Experiments with real sequences show that this method leads to accurate results and useful semantics.

show abstract

“…Higher order techniques show promise in applications for stationary signals and also for short-time transients where only a single occurrence of a signal may be available for detection ͑Dwyer, 1984; Hinich and Wilson, 1990;Kletter and Messer, 1990;Sangfelt and Persson, 1993;Delaney, 1994;Tague et al, 1994;Baugh and Hardwicke, 1994;Nuttall, 1994͒. The latter case, for correlation detectors, has been investigated in previous papers by the authors using both computer simulations ͑Pflug et al., 1992b, 1994b͒ and more recently for unknown source detection, using theoretical performance predictions for the case of uncorrelated noise ͑Pflug et al, 1995b͒.…”

Section: Introductionmentioning

confidence: 99%

Known source detection predictions for higher order correlators

Pflug

Ioup

1998

The Journal of the Acoustical Society of America

View full text Add to dashboard Cite

The problem addressed in this paper is whether higher order correlation detectors can perform better in white noise than the cross correlation detector for the detection of a known transient source signal, if additional receiver information is included in the higher order correlations. While the cross correlation is the optimal linear detector for white noise, additional receiver information in the higher order correlations makes them nonlinear. In this paper, formulas that predict the performance of higher order correlation detectors of energy signals are derived for a known source signal. Given the first through fourth order signal moments and the noise variance, the formulas predict the SNR for which the detectors achieve a probability of detection of 0.5 for any level of false alarm, when noise at each receiver is independent and identically distributed. Results show that the performance of the cross correlation, bicorrelation, and tricorrelation detectors are proportional to the second, fourth, and sixth roots of the sampling interval, respectively, but do not depend on the observation time. Also, the SNR gains of the higher order correlation detectors relative to the cross correlation detector improve with decreasing probability of false alarm. The source signal may be repeated in higher order correlations, and gain formulas are derived for these cases as well. Computer simulations with several test signals are compared to the performance predictions of the formulas. The breakdown of the assumptions for signals with too few sample points is discussed, as are limitations on the design of signals for improved higher order gain. Results indicate that in white noise it is difficult for the higher order correlation detectors in a straightforward application to achieve better performance than the cross correlation. ͓S0001-4966͑98͒01805-0͔

show abstract

Signal detection using third-order moments

Cited by 11 publications

References 35 publications

Human activity localization via sequential change detection

Human activity localization via sequential change detection

Color aided motion‐segmentation and object tracking for video sequences semantic analysis

Known source detection predictions for higher order correlators

Contact Info

Product

Resources

About