BACKGROUND

Images observed in nature or produced by instruments including biological microscopes, medical imaging systems, spatial telescopes or industrial devices often show textural aspects.

Isotropy/anisotropy is an important indicator for textural characterization and can give essential information for image treatment (segmentation or classification).

HOW IT WORKS

Marseille Institute of Mathematics offers a method that allows the automatic computation of anisotropic texture features and describes the local fluctuations of the texture regularity within digital images.

Original statistical tests allow to determine whether a texture is anisotropic or not. The programmable automatic calculator for the estimation of local Hurst index uses a new technique based on quadratic variations.

Anisotropic Brownian field texture with Hurst function ranging

KEY BENEFITS vs. STATE OF THE ART

DEVELOPMENT STATUS

Demonstrated for breast cancer risk evaluation with mammograms

Software beta version available for demonstration

APPLICATIONS

• Medical imaging (radiographic texture analysis, MRI texture classification…)
• Graphic design, meteorology, geosciences, material sciences