Classification of moving objects in atmospherically degraded video

Chen, Eli; Haik, Oren; Yitzhaky, Yitzhak

doi:10.1117/1.oe.51.10.101710

Cited by 17 publications

(18 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We used this formula to quantify the errors harvested by many works in the literature, mostly caused by a wrong extrapolation of unidimensional results without taking into account the dimensionality of the model. Using several plots, we showed that such errors can be significantly large depending on the dimension n. In particular, we discussed the deviations in the typical threshold range d ∈ [2,3].…”

Section: Discussionmentioning

confidence: 99%

“…For instance, a threshold d = 2.5 causes an error of ≈ 10% if the model has n = 3, whereas the error grows up to 40% if n = 6, and > 95% if n ≥ 10. In addition, in the typical interval d ∈ [2,3] and for some values of n, there is a wild error variation. For example, errors of 3% − 23% if n = 3, 17% − 64% if n = 6, or 50% − 90% if n = 10.…”

Section: B Case Ii: Setting the Threshold Without Taking Into Accounmentioning

confidence: 99%

“…P 1 and P 2 are represented in Figs. 3 and 4, respectively. unidimensional distribution, researchers select d ∈ [2,3]. For a small dimensionality such as n = 3, this yields moderate errors of 3% − 12%.…”

Section: A Case I: Treating Each Dimension Independentlymentioning

confidence: 99%

“…Many of these works [2] use multivariate normal distributions, or mixtures of them [3], due to the satisfactory continuity, differentiability and locality properties of the Gaussian function [4].…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

On the Mahalanobis Distance Classification Criterion for Multidimensional Normal Distributions

Gallego

Cuevas

Mohedano

et al. 2013

IEEE Trans. Signal Process.

View full text Add to dashboard Cite

Abstract-Many existing engineering works model the statistical characteristics of the entities under study as normal distributions. These models are eventually used for decision making, requiring in practice the definition of the classification region corresponding to the desired confidence level. Surprisingly enough, however, a great amount of computer vision works using multidimensional normal models leave unspecified or fail to establish correct confidence regions due to misconceptions on the features of Gaussian functions or to wrong analogies with the unidimensional case. The resulting regions incur in deviations that can be unacceptable in high-dimensional models.Here we provide a comprehensive derivation of the optimal confidence regions for multivariate normal distributions of arbitrary dimensionality. To this end, firstly we derive the condition for region optimality of general continuous multidimensional distributions, and then we apply it to the widespread case of the normal probability density function. The obtained results are used to analyze the confidence error incurred by previous works related to vision research, showing that deviations caused by wrong regions may turn into unacceptable as dimensionality increases. To support the theoretical analysis, a quantitative example in the context of moving object detection by means of background modeling is given.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: B Case Ii: Setting the Threshold Without Taking Into Accounmentioning

confidence: 99%

Section: A Case I: Treating Each Dimension Independentlymentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

On the Mahalanobis Distance Classification Criterion for Multidimensional Normal Distributions

Gallego

Cuevas

Mohedano

et al. 2013

IEEE Trans. Signal Process.

View full text Add to dashboard Cite

show abstract

“…In a previous study [4], we examined the effects of image restoration (de-blurring) on geometrical features of moving objects in long-distance imaging through atmospheric medium. Later, we examined the effect of image restoration on the classification of such moving objects [10]. In both studies, we employed a simple technique to distinguish the moving objects from background turbulence-induced movements.…”

Section: Introductionmentioning

confidence: 99%

Surveillance in long-distance turbulence-degraded videos

Yitzhaky¹,

Chen

Haik

2013

SPIE Proceedings

Self Cite

View full text Add to dashboard Cite

Surveillance in long-distance turbulence-degraded video is a difficult challenge because of the effects of the atmospheric turbulence that causes blur and random shifts in the image. As imaging distances increase, the degradation effects become more significant. This paper presents a method for surveillance in long-distance turbulence-degraded videos. This method is based on employing new criteria for discriminating true from false object detections. We employ an adaptive thresholding procedure for background subtraction, and implement new criteria for distinguishing true from false moving objects, that take into account the temporal consistency of both shape and motion properties. Results show successful detection also tracking of moving objects on challenging video sequences, which are significantly distorted with atmospheric turbulence. However, when the imaging distance is increased higher false alarms may occur. The method presented here is relatively efficient and has low complexity.

show abstract