Temporally Consistent Gaussian Random Field for Video Semantic Analysis

Gaussian random field on general ultrametric space is introduced as a solution of pseudodifferential stochastic equation. Covariation of the introduced random field is computed with the help of wavelet analysis on ultrametric spaces. Notion of ultrametric Markovianity, which describes independence of contributions to random field from different ultrametric balls is introduced. We show that the random field under investigation satisfies this property.

Download Full-text

Geometry-adapted Gaussian random field regression

2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) ◽

10.1109/icassp.2017.7953414 ◽

2017 ◽

Cited By ~ 1

Author(s):

Zhen Zhang ◽

Mianzhi Wang ◽

Yijian Xiang ◽

Arye Nehorai

Keyword(s):

Random Field ◽

Gaussian Random Field

Download Full-text

Video semantic analysis based kernel locality-sensitive discriminative sparse representation

Expert Systems with Applications ◽

10.1016/j.eswa.2018.11.016 ◽

2019 ◽

Vol 119 ◽

pp. 429-440 ◽

Cited By ~ 4

Author(s):

Ben-Bright Benuwa ◽

Yongzhao Zhan ◽

Augustine Monney ◽

Benjamin Ghansah ◽

Ernest K. Ansah

Keyword(s):

Sparse Representation ◽

Semantic Analysis ◽

Video Semantic Analysis

Download Full-text

The Transparent Dead Leaves Model

Advances in Applied Probability ◽

10.1017/s0001867800005425 ◽

2012 ◽

Vol 44 (01) ◽

pp. 1-20 ◽

Cited By ~ 1

Author(s):

B. Galerne ◽

Y. Gousseau

Keyword(s):

Central Limit Theorem ◽

Limit Theorem ◽

Random Field ◽

Central Limit ◽

Gaussian Random Field ◽

Natural Images ◽

Simulation Algorithm ◽

New Model ◽

Grain Model ◽

Dead Leaves Model

In this paper we introduce the transparent dead leaves (TDL) random field, a new germ-grain model in which the grains are combined according to a transparency principle. Informally, this model may be seen as the superposition of infinitely many semitransparent objects. It is therefore of interest in view of the modeling of natural images. Properties of this new model are established and a simulation algorithm is proposed. The main contribution of the paper is to establish a central limit theorem, showing that, when varying the transparency of the grain from opacity to total transparency, the TDL model ranges from the dead leaves model to a Gaussian random field.

Download Full-text

THE TOPOLOGY OF THE COSMIC MICROWAVE BACKGROUND ANISOTROPY ON THE SCALE ~1°

International Journal of Modern Physics D ◽

10.1142/s0218271896000229 ◽

1996 ◽

Vol 05 (04) ◽

pp. 319-362 ◽

Cited By ~ 4

Author(s):

D.I. NOVIKOV ◽

H.E. JØRGENSEN

Keyword(s):

Cosmic Microwave Background ◽

Random Field ◽

Gaussian Noise ◽

Gaussian Random Field ◽

Cosmological Models ◽

Topological Method ◽

Microwave Background ◽

Cosmic Microwave Background Anisotropy ◽

Non Gaussian

In this paper we develop the theory of clusterization of peaks in a Gaussian random field. We have obtained new mathematical results from this theory and the theory of percolation and have proposed a topological method of analysis of sky maps based on these results. We have simulated 10°×10° sky maps of the cosmic microwave background anisotropy expected from different cosmological models with 0.5°–1° resolution in order to demonstrate how this method can be used for detection of non-Gaussian noise in the maps and detection of the Doppler-peak in the spectrum of perturbation of ΔT/T.

Download Full-text