Industrial Applications of Colour Texture Classification Based on Anisotropic Diffusion

2017 ◽  
pp. 293-304 ◽  
Author(s):  
P. S. Hiremath ◽  
Rohini A. Bhusnurmath
Keyword(s):  
Anisotropic Diffusion ◽  
Texture Classification ◽  
Industrial Applications

Performance Analysis of Anisotropic Diffusion Based Colour Texture Descriptors in Industrial Applications

2017 ◽  
Vol 7 (2) ◽  
pp. 50-63 ◽  
Author(s):  
Prakash S. Hiremath ◽  
Rohini A. Bhusnurmath
Keyword(s):  
Performance Analysis ◽  
Anisotropic Diffusion ◽  
State Of The Art ◽  
Texture Classification ◽  
Industrial Applications ◽  
Statistical Features ◽  
Linear Discriminant ◽  
Class Separability ◽  
Novel Method

A novel method of colour texture analysis based on anisotropic diffusion for industrial applications is proposed and the performance analysis of colour texture descriptors is examined. The objective of the study is to explore different colour spaces for their suitability in automatic classification of certain textures in industrial applications, namely, granite tiles and wood textures, using computer vision. The directional subbands of digital image of material samples obtained using wavelet transform are subjected to anisotropic diffusion to obtain the texture components. Further, statistical features are extracted from the texture components. The linear discriminant analysis is employed to achieve class separability. The texture descriptors are evaluated on RGB, HSV, YCbCr, Lab colour spaces and compared with gray scale texture descriptors. The k-NN classifier is used for texture classification. For the experimentation, benchmark databases, namely, MondialMarmi and Parquet are considered. The experimental results are encouraging as compared to the state-of-the-art-methods.

Anisotropic Diffusion-Based Color Texture Analysis for Industrial Application

2020 ◽  
pp. 30-64
Author(s):  
Rohini A. Bhusnurmath ◽  
Prakash S. Hiremath
Keyword(s):  
Computer Vision ◽  
Texture Component ◽  
Industrial Application ◽  
Anisotropic Diffusion ◽  
Texture Classification ◽  
Industrial Applications ◽  
Color Spaces ◽  
Linear Discriminant ◽  
Benchmark Datasets ◽  
Different Color

This chapter proposes the framework for computer vision algorithm for industrial application. The proposed framework uses wavelet transform to obtain the multiresolution images. Anisotropic diffusion is employed to obtain the texture component. Various feature sets and their combinations are considered obtained from texture component. Linear discriminant analysis is employed to get the distinguished features. The k-NN classifier is used for classification. The proposed method is experimented on benchmark datasets for texture classification. Further, the method is extended to exploration of different color spaces for finding reference standard. The thrust area of industrial applications for machine intelligence in computer vision is considered. The industrial datasets, namely, MondialMarmi dataset for granite tiles and Parquet dataset for wood textures are experimented. It was observed that the combination of features performs better in YCbCr and HSV color spaces for MondialMarmi and Parquet datasets as compared to the other methods in literature.

scholarly journals Automatic Unsupervised Texture Recognition Framework Using Anisotropic Diffusion-Based Multi-Scale Analysis and Weight-Connected Graph Clustering

Symmetry ◽  
2021 ◽  
Vol 13 (6) ◽  
pp. 925
Author(s):  
Tudor Barbu
Keyword(s):  
Anisotropic Diffusion ◽  
Texture Classification ◽  
Research Work ◽  
Texture Feature ◽  
Weighted Graph ◽  
Graph Clustering ◽  
Extraction Process ◽  
Feature Vectors ◽  
Multi Scale ◽  
Similarity Graph

A novel unsupervised texture classification technique is proposed in this research work. The proposed method clusters automatically the textures of an image collection in similarity classes whose number is not a priori known. A nonlinear diffusion-based multi-scale texture analysis approach is introduced first. It creates an effective scale-space by using a well-posed anisotropic diffusion filtering model that is proposed and approximated numerically here. A feature extraction process using a bank of circularly symmetric 2D filters is applied at each scale, then a rotation-invariant texture feature vector is achieved for the current image by combining the feature vectors computed at all these scales. Next, a weighted similarity graph, whose vertices correspond to the texture feature vectors and the weights of its edges are obtained from the distances computed between these vectors, is created. A novel weighted graph clustering technique is then applied to this similarity graph, to determine the texture classes. Numerical simulations and method comparisons illustrating the effectiveness of the described framework are also discussed in this work.

Enhancing LBP by preprocessing via anisotropic diffusion

2018 ◽  
Vol 29 (08) ◽  
pp. 1850071 ◽  
Author(s):  
Mariane Barros Neiva ◽  
Patrick Guidotti ◽  
Odemir Martinez Bruno
Keyword(s):  
Texture Analysis ◽  
Significant Role ◽  
Anisotropic Diffusion ◽  
Texture Classification ◽  
Local Feature ◽  
Gaussian Kernel ◽  
Multiscale Approach ◽  
Texture Recognition ◽  
Feature Descriptors ◽  
Salient Features

The main goal of this paper is to study the addition of a new preprocessing step in order to improve local feature descriptors and texture classification. The preprocessing is implemented by using transformations which help highlight salient features that play a significant role in texture recognition. We evaluate and compare four different competing methods: three different anisotropic diffusion methods including the classical anisotropic Perona–Malik diffusion and two subsequent regularizations of it and the application of a Gaussian kernel, which is the classical multiscale approach in texture analysis. The combination of the transformed images and the original ones are analyzed. The results show that the use of the preprocessing step does lead to an improvement in texture recognition.

The Application of Ultra-Thin Sectioning Techniques to Non-Biological Samples

1968 ◽  
Vol 26 ◽  
pp. 332-333
Author(s):  
C. F. Oster
Keyword(s):  
Biological Sciences ◽  
Epoxy Resin ◽  
Cross Sections ◽  
Biological Samples ◽  
Industrial Applications ◽  
Photographic Film ◽  
Silver Particles ◽  
Thin Sectioning ◽  
Embedding Medium

Although ultra-thin sectioning techniques are widely used in the biological sciences, their applications are somewhat less popular but very useful in industrial applications. This presentation will review several specific applications where ultra-thin sectioning techniques have proven invaluable.The preparation of samples for sectioning usually involves embedding in an epoxy resin. Araldite 6005 Resin and Hardener are mixed so that the hardness of the embedding medium matches that of the sample to reduce any distortion of the sample during the sectioning process. No dehydration series are needed to prepare our usual samples for embedding, but some types require hardening and staining steps. The embedded samples are sectioned with either a prototype of a Porter-Blum Microtome or an LKB Ultrotome III. Both instruments are equipped with diamond knives.In the study of photographic film, the distribution of the developed silver particles through the layer is important to the image tone and/or scattering power. Also, the morphology of the developed silver is an important factor, and cross sections will show this structure.

The mensuration of interfaces

1992 ◽  
Vol 50 (1) ◽  
pp. 216-217
Author(s):  
W.M. Stobbs
Keyword(s):  
Electron Microscopy ◽  
50Th Anniversary ◽  
Numerical Data ◽  
Industrial Applications ◽  
Dynamical Theory ◽  
Large Strains ◽  
Reasonable Approach ◽  
Lattice Resolution ◽  
Analogue To Digital ◽  
The Way

I do not have access to the abstracts of the first meeting of EMSA but at this, the 50th Anniversary meeting of the Electron Microscopy Society of America, I have an excuse to consider the historical origins of the approaches we take to the use of electron microscopy for the characterisation of materials. I have myself been actively involved in the use of TEM for the characterisation of heterogeneities for little more than half of that period. My own view is that it was between the 3rd International Meeting at London, and the 1956 Stockholm meeting, the first of the European series , that the foundations of the approaches we now take to the characterisation of a material using the TEM were laid down. (This was 10 years before I took dynamical theory to be etched in stone.) It was at the 1956 meeting that Menter showed lattice resolution images of sodium faujasite and Hirsch, Home and Whelan showed images of dislocations in the XlVth session on “metallography and other industrial applications”. I have always incidentally been delighted by the way the latter authors misinterpreted astonishingly clear thickness fringes in a beaten (”) foil of Al as being contrast due to “large strains”, an error which they corrected with admirable rapidity as the theory developed. At the London meeting the research described covered a broad range of approaches, including many that are only now being rediscovered as worth further effort: however such is the power of “the image” to persuade that the above two papers set trends which influence, perhaps too strongly, the approaches we take now. Menter was clear that the way the planes in his image tended to be curved was associated with the imaging conditions rather than with lattice strains, and yet it now seems to be common practice to assume that the dots in an “atomic resolution image” can faithfully represent the variations in atomic spacing at a localised defect. Even when the more reasonable approach is taken of matching the image details with a computed simulation for an assumed model, the non-uniqueness of the interpreted fit seems to be rather rarely appreciated. Hirsch et al., on the other hand, made a point of using their images to get numerical data on characteristics of the specimen they examined, such as its dislocation density, which would not be expected to be influenced by uncertainties in the contrast. Nonetheless the trends were set with microscope manufacturers producing higher and higher resolution microscopes, while the blind faith of the users in the image produced as being a near directly interpretable representation of reality seems to have increased rather than been generally questioned. But if we want to test structural models we need numbers and it is the analogue to digital conversion of the information in the image which is required.

The influence of confocal microscope design on imaging performance

1993 ◽  
Vol 51 ◽  
pp. 144-145
Author(s):  
C J R Sheppard
Keyword(s):  
Image Quality ◽  
Fluorescence Imaging ◽  
Confocal Microscope ◽  
Industrial Applications ◽  
Three Dimensions ◽  
Design Parameters ◽  
Weak Signals ◽  
Size And Shape ◽  
Imaging Performance ◽  
Fundamental Limitation

The confocal microscope is now widely used in both biomedical and industrial applications for imaging, in three dimensions, objects with appreciable depth. There are now a range of different microscopes on the market, which have adopted a variety of different designs. The aim of this paper is to explore the effects on imaging performance of design parameters including the method of scanning, the type of detector, and the size and shape of the confocal aperture.It is becoming apparent that there is no such thing as an ideal confocal microscope: all systems have limitations and the best compromise depends on what the microscope is used for and how it is used. The most important compromise at present is between image quality and speed of scanning, which is particularly apparent when imaging with very weak signals. If great speed is not of importance, then the fundamental limitation for fluorescence imaging is the detection of sufficient numbers of photons before the fluorochrome bleaches.

Microstructural characterization of sol-gel coating on PET films

1994 ◽  
Vol 52 ◽  
pp. 886-887
Author(s):  
R. T. Chen ◽  
R.A. Norwood
Keyword(s):  
Ion Beam ◽  
Sol Gel ◽  
Industrial Applications ◽  
Hard Coatings ◽  
Nanometer Scale ◽  
Ion Beam Sputtering ◽  
Process Technology ◽  
Refractive Index Measurement ◽  
Organically Modified ◽  
Pet Films

Sol-gel processing has been used to control the structure of a material on a nanometer scale in preparing advanced ceramics and glasses. Film coating using the sol-gel process was also found to be a viable process technology in applications such as optical, porous, antireflection and hard coatings. In this study, organically modified silicate (Ormosil) coatings are applied to PET films for various industrial applications. Sol-gel materials are known to exhibit nanometer scale structures which havepreviously been characterized by small-angle X-ray scattering (SAXS), neutron scattering and light scattering. Imaging of the ultrafine sol-gel structures has also been performed using an ultrahigh resolution replica/TEM technique. The objective of this study was to evaluate the ultrafine structures inthe sol gel coatings using a direct imaging technique: atomic force microscopy (AFM). In addition, correlation of microstructures with processing parameters, coating density and other physical properties will be discussed.The materials evaluated are organically modified silicate coatings on PET film substrates. Refractive index measurement by the prism coupling method was used to assess density of the sol-gel coating.AFM imaging was performed on a Nanoscope III AFM (by Digital Instruments) using constant force mode. Solgel coating samples coated with a thin layer of Ft (by ion beam sputtering) were also examined by STM in order to confirm the structures observed in the contact type AFM. In addition, to compare the previous results, sol-gel powder samples were also prepared by ultrasonication followed by Pt/Au shadowing and examined using a JEOL 100CX TEM.

Sign in / Sign up

Export Citation Format

Share Document