Volumes of Blurred-Invariant Gaussians for Dynamic Texture Classification

Abstract : An effective model, which jointly captures shape and motion cues, for dynamic texture (DT) description is introduced by taking into account advantages of volumes of blurred-invariant features in three main following stages. First, a 3-dimensional Gaussian kernel is used to form smoothed sequences that allow to deal with well-known limitations of local encoding such as near uniform regions and sensitivity to noise. Second , a receptive volume of the Difference of Gaussians (DoG) is figured out to mitigate the negative impacts of environmental and illumination changes which are major challenges in DT understanding. Finally, a local encoding operator is addressed to construct a discriminative descriptor of enhancing patterns extracted from the filtered volumes. Evaluations on benchmark datasets (i.e., UCLA, DynTex, and DynTex++) for issue of DT classification have positively validated our crucial contributions.
Complete list of metadatas

Cited literature [41 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-02160704
Contributor : Thanh Phuong Nguyen <>
Submitted on : Wednesday, June 19, 2019 - 10:29:23 PM
Last modification on : Friday, June 21, 2019 - 11:16:33 AM

File

CAIP-34.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-02160704, version 1

Citation

Thanh Tuan Nguyen, Thanh Phuong Nguyen, Frédéric Bouchara, Ngoc-Son Vu. Volumes of Blurred-Invariant Gaussians for Dynamic Texture Classification. Computer Analysis of Images and Patterns (CAIP), Sep 2019, Salerno, Italy. ⟨hal-02160704⟩

Share

Metrics

Record views

86

Files downloads

26