FWLBP: A Scale Invariant Descriptor for Texture Classification (original) (raw)

Locally Invariant Fractal Features for Statistical Texture Classification

2007 IEEE 11th International Conference on Computer Vision, 2007

We address the problem of developing discriminative, yet invariant, features for texture classification. Texture variations due to changes in scale are amongst the hardest to handle. One of the most successful methods of dealing with such variations is based on choosing interest points and selecting their characteristic scales [Lazebnik et al. PAMI 2005]. However, selecting a characteristic scale can be unstable for many textures. Furthermore, the reliance on an interest point detector and the inability to evaluate features densely can be serious limitations. Fractals present a mathematically well founded alternative to dealing with the problem of scale. However, they have not become popular as texture features due to their lack of discriminative power. This is primarily because: (a) fractal based classification methods have avoided statistical characterisations of textures (which is essential for accurate analysis) by using global features; and (b) fractal dimension features are unable to distinguish between key texture primitives such as edges, corners and uniform regions. In this paper, we overcome these drawbacks and develop local fractal features that are evaluated densely. The features are robust as they do not depend on choosing interest points or characteristic scales. Furthermore, it is shown that the local fractal dimension is invariant to local bi-Lipschitz transformations whereas its extension is able to correctly distinguish between fundamental texture primitives. Textures are characterised statistically by modelling the full joint PDF of these features. This allows us to develop a texture classification framework which is discriminative, robust and achieves state-of-the-art performance as compared to affine invariant and fractal based methods.

A statistical descriptor for texture images based on the box counting fractal dimension

Physica A: Statistical Mechanics and its Applications, 2019

New texture image descriptors are proposed. • The descriptors are based on a statistical analysis of the box counting fractal dimension. • The descriptors are tested in the classification of UIUC, Outex and USPTex databases. • Other state-of-the-art approaches are outperformed in texture classification. • A formal mathematical analysis of the model is also presented.

Fractal signatures for complex natural textures recognition

IECON '98. Proceedings of the 24th Annual Conference of the IEEE Industrial Electronics Society (Cat. No.98CH36200), 1998

Random textures can be considered fractal sets under certain circumstances. The main parameter of a fractal set is its dimension. The description of the fractal dimension over the spatial scale is called fractal signature. Some methods for defining the fractal dimension are described, with special emphasis to the Hausdorff and box counting. Fractal signatures are defined based on two viewpoints: upside view, from the brightest pixels, and downside view, for the darkest ones. Signatures' geometry is compared for different textures. Natural images are taken as fractals, and their signatures evaluated as features for texture recognition. Cork-agglomerated images with defects, and patches of office and CRAFT paper images were taken as fractal sets and their signatures were analysed. Results were discussed and some conclusions were draft.

Texture Classification Using Fractal Dimension and Lacunarity

2014

The fractal features including fractal dimension (FD) and lacunarity measures are often used as indicators of texture. Several FD and lacunarity estimation methods leading to different results have been proposed in the literature. This paper is devoted mainly to show the need to combine the lacunarity with fractal dimension for the discrimination between different textures and especially to check if this combination is valid with any FD estimation method. Keywordstexture analysis; fractal dimension; lacunarity; classification.

Texture descriptor combining fractal dimension and artificial crawlers

Physica A: Statistical Mechanics and its Applications, 2014

Texture is an important visual attribute used to describe images. There are many methods available for texture analysis. However, they do not capture the details richness of the image surface. In this paper, we propose a new method to describe textures using the artificial crawler model. This model assumes that each agent can interact with the environment and each other. Since this swarm system alone does not achieve a good discrimination, we developed a new method to increase the discriminatory power of artificial crawlers, together with the fractal dimension theory. Here, we estimated the fractal dimension by the Bouligand-Minkowski method due to its precision in quantifying structural properties of images. We validate our method on two texture datasets and the experimental results reveal that our method leads to highly discriminative textural features. The results indicate that our method can be used in different texture applications.

Color texture analysis based on fractal descriptors

Pattern Recognition, 2012

Color texture classification is an important step in image segmentation and recognition. The color information is especially important in textures of natural scenes, such as leaves surfaces, terrains models, etc. In this paper, we propose a novel approach based on the fractal dimension for color texture analysis. The proposed approach investigates the complexity in R, G and B color channels to characterize a texture sample. We also propose to study all channels in combination, taking into consideration the correlations between them. Both these approaches use the volumetric version of the Bouligand-Minkowski Fractal Dimension method. The results show a advantage of the proposed method over other color texture analysis methods.

Scale selective extended local binary pattern for texture classification

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

In this paper, we propose a new texture descriptor, scale selective extended local binary pattern (SSELBP), to characterize texture images with scale variations. We first utilize multiscale extended local binary patterns (ELBP) with rotationinvariant and uniform mappings to capture robust local microand macro-features. Then, we build a scale space using Gaussian filters and calculate the histogram of multi-scale ELBPs for the image at each scale. Finally, we select the maximum values from the corresponding bins of multi-scale ELBP histograms at different scales as scale-invariant features. A comprehensive evaluation on public texture databases (KTH-TIPS and UMD) shows that the proposed SSELBP has high accuracy comparable to state-of-the-art texture descriptors on gray-scale-, rotation-, and scale-invariant texture classification but uses only one-third of the feature dimension.

Texture features based on an efficient local binary pattern descriptor

Computers & Electrical Engineering, 2018

Texture characterization aims at describing the spatial arrangement of local structures within an image. However, mixed pixels that are generally located near boundaries of the regions represent challenge to perform accurate image texture discrimination. To address this problem, this paper proposes a robust discriminating texture features relying on an efficient Local Binary Pattern (LBP) descriptor, where the spatial information within image is taken into account. To determine for each pixel both a proper scale parameter and a threshold value to compute the LBP code, an efficient way relying on bilateral filter-based multi-scale image analysis is used. First, the difference of Gaussian operator is used to determine the corresponding scale. Second, key points based-approach is used to identify the threshold value of each pixel. This provides the ability to deal with mixed pixels. Then, LBP code is computed to characterize the texture information for each pixel. Experimental results, using both synthetic and real images, show that the proposed appropriate-scalethreshold selection strategy demonstrates a significant improvement in texture discrimination ability.

Texture Recognition using Hybrid Fractal and Blocking Approach

International Journal of Computer Applications, 2014

Texture analysis is an important and useful area of study in machine vision. Most natural surfaces exhibit texture and a successful vision system must be able to deal with such like surfaces. Many natural surfaces have a statistical quality of roughness and self-similarity at different scales. Fractals are very useful and have become popular in modeling these properties in image processing. This work adopts analyzing samples by three methods fractal dimension, block approach and Hybrid method (fractal dimension method with block approach). The fractal dimension get a highest recognition rate among remaining used methods, it obtain a rate 95% as compare with 40% Block Approach model, 65.5% Hybrid method. The results show the efficiency of fractal dimension recognition than blocking approach recognition and hybrid recognition in textures.

Texture Classification Using Spline, Wavelet Decomposition and Fractal Dimension

Applied and Computational Mathematics, 2015

Feature extraction is an important process for texture classification. This paper suggests two sets of features for texture analysis. In the first set of features, a set of fractal features is obtained from the eight wavelet sub-bands that are generated by applying Haar wavelet transform twice times according to dyadic architecture. The fractal features are determined using the differential box counting method. While for determining the second set of features, the cubic spline representation is applied to decompose the image signal into rough and smooth components; then applying the wavelet transform and finally compute the fractal dimension for all the sub-bands of both images. Each type of these two extracted feature sets is studied individually, and they are used together. Their overall performance is investigated. The proposed features set has been applied on two texture datasets, one consists of textures with directional properties, and the second set consists of textures samples that have directional attributes. The test results showed that the proposed methods give a high level of classification with images that have or do not have directional properties.