Modeling the Perceptual Similarity of Static and Dynamic Visual Textures -Application to the Perceptual Optimization of Video Compression

Karam Naser 1, 2
Abstract : Textures are special signals in the visual scene, where they can cover large areas. They can be classified into two categories: static and dynamic, where dynamic textures involve temporal variations. Several works on the perception of static textures made it possible to define visual similarity measurements for applications such as the recognition or classification of textures. These measures often use a representation inspired by the neural processing of the human visual system. However, such approaches have been little explored in the case of dynamic textures. In this thesis, a generalized perceptual model for the measurement of similarity applicable to static and dynamic textures has been developed. This model is inspired by the processing performed in the primary visual cortex. It is very effective for texture classification and recognition applications. The application of the model in the context of the perceptual optimization of video compression, was also studied. In particular, the integration of the similarity measure between textures, was used for the rate-distortion optimization of the encoder. Experimental results with human observers showed an improved visual quality of the decoded videos, with a significant reduction in the bitrate compared to the traditional approaches.
Complete list of metadatas

Cited literature [209 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-01653260
Contributor : Vincent Ricordel <>
Submitted on : Friday, December 1, 2017 - 11:35:11 AM
Last modification on : Tuesday, March 26, 2019 - 9:25:22 AM

File

these_KaramNaser-FINAL.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-01653260, version 1

Collections

Citation

Karam Naser. Modeling the Perceptual Similarity of Static and Dynamic Visual Textures -Application to the Perceptual Optimization of Video Compression. Image Processing [eess.IV]. Université Bretagne Loire; Université de Nantes, 2017. English. ⟨tel-01653260⟩

Share

Metrics

Record views

361

Files downloads

117