Elsevier

Neurocomputing

Volume 149, Part C, 3 February 2015, Pages 1560-1572
Neurocomputing

Improved texture image classification through the use of a corrosion-inspired cellular automaton

https://doi.org/10.1016/j.neucom.2014.08.036Get rights and content

Highlights

  • A cellular automata (CA) approach for texture analysis and recognition is proposed.

  • The models were inspired in the corrosion model of alloy surfaces.

  • The image is modeled as an alloy surface and the CA simulates the corrosion evolution.

  • The corrosion behavior provides features that are used to characterize the texture.

  • Results show the proposed method outperformed state-of-the-art methods.

Abstract

In this paper, the problem of classifying synthetic and natural texture images is addressed. To tackle this problem, an innovative method is proposed that combines concepts from corrosion modeling and cellular automata to generate a texture descriptor. The core processes of metal (pitting) corrosion are identified and applied to texture images by incorporating the basic mechanisms of corrosion in the transition function of the cellular automaton. The surface morphology of the image is analyzed before and during the application of the transition function of the cellular automaton. In each iteration the cumulative mass of corroded product is obtained to construct each of the attributes of the texture descriptor. In the final step, this texture descriptor is used for image classification by applying Linear Discriminant Analysis. The method was tested on the well-known Brodatz and Vistex databases. In addition, in order to verify the robustness of the method, its invariance to noise and rotation was tested. To that end, different variants of the original two databases were obtained through addition of noise to and rotation of the images. The results showed that the proposed texture descriptor is effective for texture classification according to the high success rates obtained in all cases. This indicates the potential of employing methods taking inspiration from natural phenomena in other fields.

Introduction

The classification of texture images is an important problem in pattern recognition and consequently forms the subject of many research works in this field. Texture is an important image feature with a strong discriminative capability and is therefore widely used in computer vision. Image descriptors for image texture are obtained from the analysis of groups of pixels and the way this analysis is performed is used to classify the different methods of texture analysis. Based on the domain from which the texture feature is extracted, five main categories can be distinguished: structural [1], [2], [3], statistical [4], model-based [5], [6], [7], spectral [8], [9], and agent-based methods [10], [11], [12].

This paper proposes a novel texture descriptor constructed by means of a cellular automaton (CA) taking inspiration from the pitting corrosion phenomenon, further on referred to as the Corrosion-Inspired Texture Analysis (CITA) descriptor. The basic mechanisms behind this detrimental reaction which occurs between metals (or alloys) and their environment serve as inspiration to develop a CA-based model. Next, this CA-based model is employed to generate a texture descriptor for classification by treating the image to be classified as a metal surface. The CA-based model, like real corrosion, amplifies existing differences in material and height (in this case grayscale value) so that the biggest contrasts in the original texture image will become more pronounced and smaller contrasts will be nullified. The eroded mass of ‘metal’ by the progression of pitting corrosion at each iteration is used to generate a texture descriptor that describes the image to be classified. These texture descriptors are then used as feature vectors in a supervised setting to develop a classification method. The effectiveness of this strategy is demonstrated on two texture databases, Brodatz and Vistex, with natural and synthetic textures. In addition, to verify the robustness of the classification method, its invariance to noise and rotation were tested, obtaining satisfactory results.

The main contribution of this work thus lies in demonstrating that a natural phenomenon can be a source of inspiration to develop a robust texture descriptor for the classification of both natural and synthetic texture images. Moreover, the proposed method outperforms the state-of-the-art methods in texture analysis, thus contributing to the image analysis field. Our paper is organized as follows. Section 2 describes the basics behind the pitting corrosion phenomenon, while the definition of a CA as well as further explanation of some parts of this definition form the subject of Section 3. The classification method is described in Section 4 and the experimental setup needed to test its efficacy is explained in Section 5. Section 6 presents the results and Section 7 presents the discussion of the study. Finally, the paper is concluded in Section 8.

Section snippets

Pitting corrosion

Corrosion is the disintegration of metals (and alloys) into their constituents due to reaction with the environment and is one of the main causes of structural failure in industrial systems, and poses as such an economic problem [13]. Dealing with corrosion is difficult because of its complex nature and the involvement of many variables. Therefore, modeling and simulation could allow for predicting more accurately the corrosion process in time. CA-based models are excellent candidates for

Cellular automata

CAs are mathematical constructs in which the space, state and time domains are discrete as opposed to partial differential equations (PDE) in which these three domains are continuous [23], [24]. The ability of CAs to generate a rich spectrum of sometimes complex spatio-temporal patterns from relatively simple underlying transition functions has led to their successful employment in the modelling of several biological processes [25], [26], [27], [28], [29], [30]. Models based on CAs can be seen

Corrosion-inspired texture analysis

To obtain the texture descriptor proposed in this paper, initially, the texture image is converted into the initial state of a CA. Thereafter, a CA-based model inspired by the pitting corrosion phenomenon is evaluated for a number of time steps. The cumulative mass of corroded metal after each iteration of the CA-based model is used to construct a texture descriptor for every texture image. Finally, these texture descriptors are used to classify the images via Linear Discriminant Analysis

Experimental setup

To investigate the performance of the classification method, it is employed for the classification of the images of two classical texture databases, the Brodatz and Vistex databases, and the results are compared to those obtained with several established features from the literature. The remainder of this section includes a short description of the employed databases and the features from the literature used for comparison as well as an optimization of the parameters of the procedure to obtain

Results

This section reports on the performance of the proposed classification method. The classification performance when using the CITA descriptor is compared to that when using the traditional texture features in the literature. Three sets of experiments were performed: firstly on the original test databases and subsequently on modified versions of the test databases to test noise and rotation invariance. All tests were performed using the optimized values for ν, γ and the number of iterations

Discussion

Our findings suggest that a CA-based model, taking inspiration from the natural phenomenon of pitting corrosion, results in informative features for subsequent texture classification. Experiments using the original Brodatz and Vistex databases have shown the capability of the CITA descriptor to discriminate between different textural classes. However, for the Vistex database, GLDM achieves the same success rate. The latter can be explained by the fact that, just like the CITA descriptor, GLDM

Conclusions

In this paper, a new descriptor for texture analysis was proposed by combining concepts from corrosion engineering, cellular automata and pattern recognition. The developed CITA descriptor in combination with LDA was used to classify the texture images of two well-known databases: Brodatz and Vistex. The descriptor was derived from images of the original databases and the robustness of the classification method under addition of noise and rotation was investigated. For this purpose, several new

Acknowledgments

Núbia Rosa da Silva acknowledges support from FAPESP (The State of São Paulo Research Foundation), Grant no. 2011/21467-9. Pieter Van der Weeën was sponsored by the Fund for Scientific Research in Flanders (FWO). Odemir Martinez Bruno gratefully acknowledges the financial support of CNPq (National Council for Scientific and Technological Development, Brazil) (Grant nos. 308449/2010-0 and 473893/2010-0) and FAPESP (Grant no. 2011/01523-1).

Núbia Rosa da Silva received the BA degree in Computer Science from Federal University of Goiás in 2007 and the M.Sc. degree in Computer Science from Federal University of Uberlândia in 2010. Now, she is a Ph.D. candidate in Computer Science at University of São Paulo. She joined the Scientific Computing Group in 2010 and her research interests include computer vision, image processing, and pattern recognition.

References (46)

  • T. Toffoli

    Cellular automata as an alternative to (rather than an approximation of) differential equations in modeling physics

    Physica D

    (1984)
  • J.M. Baetens et al.

    Effects of asynchronous updating in cellular automata

    Chaos, Solitons Fractals

    (2012)
  • A.R. Backes et al.

    Color texture analysis based on fractal descriptors

    Pattern Recognit.

    (2012)
  • M. Idrissa et al.

    Texture classification using Gabor filters

    Pattern Recognit. Lett.

    (2002)
  • Z. Guo et al.

    Rotation invariant texture classification using lbp variance (lbpv) with global matching

    Pattern Recognit.

    (2010)
  • R. Goyal, W. Goh, D. Mital, K. Chan, Scale and rotation invariant texture analysis based on structural property, in:...
  • J. Serra

    Image Analysis and Mathematical Morphology

    (1983)
  • R.M. Haralick

    Statistical and structural approaches to texture

    Proc. IEEE

    (1979)
  • A.R. Backes et al.

    Plant leaf identification based on volumetric fractal dimension

    Int. J. Pattern Recognit. Artif. Intell.

    (2009)
  • A.R. Backes et al.

    A new approach to estimate fractal dimension of texture images

  • F. Cohen et al.

    Classification of rotated and scaled textured images using Gaussian Markov random field models

    IEEE Trans. Pattern Anal. Mach. Intell.

    (1991)
  • C. Chen, A Comparative Study of Texture Classification Using Spectral Features, Technical Report ADA109408, pp....
  • X. Tang, W. Stewart, Texture classification using wavelet packet and Fourier transforms, in: Proceedings of Challenges...
  • Cited by (26)

    • A cellular automata approach to local patterns for texture recognition

      2021, Expert Systems with Applications
      Citation Excerpt :

      Since its popularization in the sixties, chaos theory, in its modern sense, has attracted attention in numerous areas. Models for image processing and analysis have also been proposed, benefiting from tools originally developed for the analysis of chaotic systems (Gao & Yang, 2014; Rosin, 2010; Guo, Wu, Ju, Wang, & Zhao, 2010; da Silva, der Ween, Baets, & Bruno, 2015; Yu, 2017). Among the classical models of deterministic chaos, cellular automata (CA) (Wolfram, 2002) have been natural candidates to model digital images, mainly due to their intrinsic representation over a two-dimensional grid, which allows the association between pixel intensity and cell states.

    • Fusion of complex networks and randomized neural networks for texture analysis

      2020, Pattern Recognition
      Citation Excerpt :

      An example is the set of techniques that use learning [55], such as descriptors based on a vocabulary of scale invariant feature transform (SIFT) [11], often called bag-of-visual-words (BOVW). Methods based on image complexity analysis are also gaining attention such as cellular automata [13] and complex networks (CN) [3,8,50]. In particular, methods based on the CN theory have achieved promising results due to their capacity to represent the relationships among structural elements of texture.

    • Multilayer complex network descriptors for color–texture characterization

      2019, Information Sciences
      Citation Excerpt :

      More recent and innovative techniques are approaching texture differently. For instance, there are texture descriptors based on image complexity analysis with fractal dimension [44], Complex Networks (CNs) [5,18,40,48] and Cellular Automaton [15]. Learning techniques also received attention in the past years, for instance, texture properties learned with a vocabulary of Scale Invariant Feature Transform (SIFT) [14], which then inspired subsequent methods.

    • An edge detection method using outer Totalistic Cellular Automata

      2016, Neurocomputing
      Citation Excerpt :

      The most characteristic feature of CA is that using simple rules interacting with a local neighbourhood it can produce very complex behaviour. Many researchers have investigated the possibility of using Cellular Automata for image processing [12,13], and a few focused on edge detection with either binary, greyscale or colour images as inputs. The linear set of rules applied to binary images were recently investigated by several authors.

    • Principal curvatures based rotation invariant algorithms for efficient texture classification

      2016, Neurocomputing
      Citation Excerpt :

      Discrete wavelet transform (DWT) [9], complex wavelet transform (CWT) [10], discrete Fourier transform (DFT) [11] and cellular automaton (CA) [12] are extensively used in texture examination. To tackle the problem of classifying natural and synthetic texture images, a robust approach that combines concepts from CA and corrosion modeling to obtain a descriptor vector has been suggested [12]. Li et al. [13] recommended a rotation invariant feature descriptor which is robust to illumination and noise by utilizing Rapid-transform.

    View all citing articles on Scopus

    Núbia Rosa da Silva received the BA degree in Computer Science from Federal University of Goiás in 2007 and the M.Sc. degree in Computer Science from Federal University of Uberlândia in 2010. Now, she is a Ph.D. candidate in Computer Science at University of São Paulo. She joined the Scientific Computing Group in 2010 and her research interests include computer vision, image processing, and pattern recognition.

    Pieter Van der Weeën (1987) holds an M.Sc. in Bioscience Engineering (2010), and a Ph.D. in Applied Biological Sciences (2014) both from Ghent University (Belgium).

    Bernard De Baets (1966) holds an M.Sc. in Maths (1988), a Postgraduate degree in Knowledge Technology (1991) and a Ph.D. in Maths (1995), all summa cum laude from Ghent University (Belgium) and is a Government of Canada Award holder (1988). He is a Full Professor in Applied Maths (1999) at Ghent University, where he is leading KERMIT, the research unit Knowledge-Based Systems. He is an Honorary Professor of Budapest Tech (2006) and an IFSA Fellow (2011). His publications comprise more than 350 papers in international journals and about 60 book chapters. He serves on the Editorial Boards of various international journals, in particular as co-editor-in-chief of Fuzzy Sets and Systems. B. De Baets coordinates EUROFUSE, the EURO Working Group on Fuzzy Sets, and is a member of the Board of Directors of EUSFLAT and of the Administrative Board of the Belgian OR Society.

    Odemir Martinez Bruno is an Associate Professor at the Physics Institute of S. Carlos at the University of S. Paulo in Brazil and head of the Scientific Computer Group. He received his B.Sc. in Computer Science (1992), from the Piracicaba Engineering College (Brazil), his M.Sc. in Applied physics (1995) and his Ph.D. in Computational Physics (2000) at the University of S. Paulo (Brazil). His fields of interest include Computer Vision, Image Analysis, Chaos, Fractals, Computational Physics, Pattern Recognition and Bioinformatics. He is an author of many papers (journal and proceedings) and several book chapters. He is a co-author of two books (Optical and Physiology of Vision: a multidisciplinary approach (Portuguese only) and Internet programming with PHP (Portuguese only)) and an inventor of seven patents.

    View full text