Francesco Bianconi

Evaluation of the effects of Gabor filter parameters on texture classification

Gabor filtering is a widely adopted technique for texture analysis. The design of a Gabor filter bank is a complex task. In texture classification, in particular, Gabor filters show a strong dependence on a certain number of parameters, the values of which may significantly affect the outcome of the classification procedures. Many different approaches to Gabor filter design, based on mathematical and physiological consideration, are documented in literature. However, the effect of each parameter, as well as the effects of their interaction, remain unclear. The overall aim of this work is to investigate the effects of Gabor filter parameters on texture classification. An extensive experimental campaign has been conducted. The outcomes of the experimental activity show a significant dependence of the percentage of correct classification on the smoothing parameter of the Gabor filters. On the contrary, the correlation between the number of frequencies and orientations used to define a filter bank and the percentage of correct classification appeared to be poor.

Evaluation of robustness against rotation of LBP, CCR and ILBP features in granite texture classification

The aim of the paper is to conduct a performance evaluation where several texture descriptors such as Local Binary Patterns (LBP), Coordinated Clusters Representation (CCR) and (Improved Local Binary Patterns) ILBP are applied for granite texture classification. In our work we were particularly interested to assess the robustness of the analysed texture descriptors to image rotation when they were implemented in both the standard and rotation-invariant forms. In order to attain this goal, we have generated a database of granite textures that were rotated using hardware and software procedures. The experimental data indicate that the ILBP features return improved performance when compared with those achieved by the LBP and CCR descriptors. Another important finding resulting from this investigation reveals that the classification results obtained when the texture analysis techniques were applied to granite image data rotated by software procedures are inconsistent with those achieved when the hardware-rotated data are used for classification purposes. This discovery is surprising and suggests that the results obtained when the texture analysis techniques are evaluated on synthetically rotated data need to be interpreted with care, as the principal characteristics of the texture are altered by the data interpolation that is applied during the image rotation process.

Theoretical and experimental comparison of different approaches for color texture classification

Color texture classification has been an area of intensive research activity. From the very onset, approaches to combining color and texture have been the subject of much discussion, and in particular, whether they should be considered joint or separately. We present a comprehensive comparison of the most prominent approaches both from a theoretical and experimental standpoint. The main contributions of our work are: (i) the establishment of a generic and extensible framework to classify methods for color texture classification on a mathematical basis, and (ii) a theoretical and experimental comparison of the most salient existing methods. Starting from an extensive set of experiments based on the Outex dataset, we highlight those texture descriptors that provide good accuracy along with low dimensionality. The results suggest that separate color and texture processing is the best practice when one seeks for optimal compromise between accuracy and limited number of features. We believe that our work may serve as a guide for those who need to choose the appropriate method for a specific application, as well as a basis for the development of new methods.

Rotation-invariant colour texture classification through multilayer CCR

The Coordinated Clusters Representation (CCR) is a texture descriptor based on the probability of occurrence of elementary binary patterns (texels) defined over a square window. The CCR was originally proposed for binary textures, and it was later extended to grayscale texture images through global image thresholding. The required global binarization is a critical point of the method, since this preprocessing stage can wipe out textural information. Another important drawback of the original CCR model is its sensitivity against rotation. In this paper we present a rotation-invariant CCR-based model for colour textures which yields a twofold improvement over the grayscale CCR: first, the use of rotation-invariant texels makes the model insensitive against rotation; secondly, the new texture model benefits from colour information and does not need global thresholding. The basic idea of the method is to describe the textural and colour content of an image by splitting the original colour image into a stack of binary images, each one representing a colour of a predefined palette. The binary layers are characterized by the probability of occurrence of rotation-invariant texels, and the overall feature vector is obtained by concatenating the histograms computed for each layer. In order to quantitatively assess our approach, we performed experiments over two datasets of colour texture images using five different colour spaces. The obtained results show robust invariance against rotation and a marked increase in classification accuracy with respect to grayscale versions of CCR and LBP.

Automatic classification of granite tiles through colour and texture features

This paper is about the development of an expert system for automatic classification of granite tiles through computer vision. We discuss issues and possible solutions related to image acquisition, robustness against noise factors, extraction of visual features and classification, with particular focus on the last two. In the experiments we compare the performance of different visual features and classifiers over a set of 12 granite classes. The results show that classification based on colour and texture is highly effective and outperforms previous methods based on textural features alone. As for the classifiers, Support Vector Machines show to be superior to the others, provided that the governing parameters are tuned properly.

An approach for feature semantics recognition in geometric models

Abstract This paper describes a method for the recognition of the semantics of parts (features) of a component from a pure geometric representation. It is suitable for verifying product life-cycle requirements from the early stages of the design process. The proposed method is appropriate to analyse B-rep geometric models, and it is not limited to models described by planar and cylindrical surfaces, but it can handle several types of face shapes. In this work the concept of semanteme is introduced. A semanteme represents the minimal element of engineering meaning that can be recognised in a geometric model. The semantemes recognised in a part of the model, which are potentially of engineering significance, are used to associate an engineering meaning to the part. This approach gives a wide flexibility to the proposed system, which is suitable to be used in different contexts of application, since it is possible to describe the reference context using the semanteme that the system can manage. In the paper the implemented prototype system is briefly described. The prototype system takes advantage of neutral interfaces that allow geometrical and topological information to be retrieved from a commercial CAD system.

Multi-class texture analysis in colorectal cancer histology.

Automatic recognition of different tissue types in histological images is an essential part in the digital pathology toolbox. Texture analysis is commonly used to address this problem; mainly in the context of estimating the tumour/stroma ratio on histological samples. However, although histological images typically contain more than two tissue types, only few studies have addressed the multi-class problem. For colorectal cancer, one of the most prevalent tumour types, there are in fact no published results on multiclass texture separation. In this paper we present a new dataset of 5,000 histological images of human colorectal cancer including eight different types of tissue. We used this set to assess the classification performance of a wide range of texture descriptors and classifiers. As a result, we found an optimal classification strategy that markedly outperformed traditional methods, improving the state of the art for tumour-stroma separation from 96.9% to 98.6% accuracy and setting a new standard for multiclass tissue separation (87.4% accuracy for eight classes). We make our dataset of histological images publicly available under a Creative Commons license and encourage other researchers to use it as a benchmark for their studies.

Performance analysis of colour descriptors for parquet sorting

In this paper we consider the problem of colour-based sorting hardwood parquet slabs into lots of similar visual appearance. As a basis for the development of an expert system to perform this task, we experimentally investigate and compare the performance of various colour descriptors (i.e.: soft descriptors, percentiles, marginal histograms and 3D histogram), and colour spaces (i.e.: RGB, HSV and CIE Lab). The results show that simple and compact colour descriptors, such as the mean of each colour channel, are as accurate as more complicated features. Likewise, we found no statistically significant difference in the accuracy attainable through the colour spaces considered in the paper. Our experiments also show that most methods are fast enough for real-time processing. The results suggest the use of simple statistical descriptors along with RGB data as the best practice to approach the problem.

Object-Based Greenhouse Classification from GeoEye-1 and WorldView-2 Stereo Imagery

Remote sensing technologies have been commonly used to perform greenhouse detection and mapping. In this research, stereo pairs acquired by very high-resolution optical satellites GeoEye-1 (GE1) and WorldView-2 (WV2) have been utilized to carry out the land cover classification of an agricultural area through an object-based image analysis approach, paying special attention to greenhouses extraction. The main novelty of this work lies in the joint use of single-source stereo-photogrammetrically derived heights and multispectral information from both panchromatic and pan-sharpened orthoimages. The main features tested in this research can be grouped into different categories, such as basic spectral information, elevation data (normalized digital surface model; nDSM), band indexes and ratios, texture and shape geometry. Furthermore, spectral information was based on both single orthoimages and multiangle orthoimages. The overall accuracy attained by applying nearest neighbor and support vector machine classifiers to the four multispectral bands of GE1 were very similar to those computed from WV2, for either four or eight multispectral bands. Height data, in the form of nDSM, were the most important feature for greenhouse classification. The best overall accuracy values were close to 90%, and they were not improved by using multiangle orthoimages.

On the Occurrence Probability of Local Binary Patterns: A Theoretical Study

It is well-known that local binary pattern (LBP) histograms of real textures exhibit a markedly uneven distribution, which is dominated by the so-called uniform patterns. The widely accepted interpretation of this phenomenon is that uniform patterns correspond to texture microfeatures, such as edges, corners, and spots. In this paper we present a theoretical study about the relative occurrence of LBPs based on the consideration that the LBP operator partitions the set of grayscale patterns into an ensemble of disjoint multidimensional polytopes. We derive exact prior probabilities of LBPs by calculating the volume of such polytopes. Our study puts in evidence that both the uneven distribution of the LBP histogram and the high occurrence of uniform patterns are direct consequences of the mathematical structure of the method rather than an intrinsic property of real textures.