Local Binary Patterns and Extreme Learning Machine based Texture Classification of Marbles

Turan, Erhan; Uçar, Ferhat; Dandil, Besir

doi:10.1109/idap.2018.8620925

Cited by 4 publications

(5 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Local Binary Pattern is a typical feature descriptor with high power discriminator and low computational cost. This fact explains why it has been used in many applications for digital image processing such as texture classification (TURAN et al, 2018), hands gestures classification (Tabatabaei and Chalechale, 2018) and face recognition (Salyut and Kurnaz, 2018).…”

Section: Local Binary Patternmentioning

confidence: 99%

CMLBPIncoherent: a New Contextual Image Descriptor for Scene Classification

Ribeiro

Salles

2019

Anais Do 14º Simpósio Brasileiro De Automação Inteligente

View full text Add to dashboard Cite

Scene classification has been widely studied in the last years and is a significant computer vision problem. Currently, this application has been involved by several researchers, since the creation of robust datasets until the utilization of high-level techniques with great performance. However, due to the high complexity and variability from scenes, there is still so much to be studied and implemented aiming to improve better results. This paper proposes a new visual descriptor for scene classification, the CMLBPIncoherent, which models the distribution of local structures through Local Binary Pattern (LBP) algorithm combined with contextual information, and discards homogeneous regions by Color Coherent Vector (CCV) algorithm. The proposal was tested using some traditional scene databases and the results illustrate a better performance in comparison with other visual descriptors known in the literature.

show abstract

Section: Local Binary Patternmentioning

confidence: 99%

CMLBPIncoherent: a New Contextual Image Descriptor for Scene Classification

Ribeiro

Salles

2019

Anais Do 14º Simpósio Brasileiro De Automação Inteligente

View full text Add to dashboard Cite

show abstract

“…Düşük hafıza tüketimi ve hızlı hesaplama özelliğinden dolayı LBP özellik çıkarım yöntemi mermer, granit, ahşap veya tekstil malzemelerinin sınıflandırılmasında kullanılmaktadır [4][5][6]. Ayrıca LBP özellik çıkarımı kullanılarak aynı türdeki mermerleri seleksiyonlarına ayıran bir çalışma gerçekleştirilmiştir [7]. Bu çalışmada, ELM sınıflandırıcısıyla birlikte LBP, SIFT özellik çıkarım yöntemine göre daha başarımı yüksek sonuçlar üretebildiği gösterilmiştir [7].…”

Section: Gi̇ri̇ş (Introduction)unclassified

“…Ayrıca LBP özellik çıkarımı kullanılarak aynı türdeki mermerleri seleksiyonlarına ayıran bir çalışma gerçekleştirilmiştir [7]. Bu çalışmada, ELM sınıflandırıcısıyla birlikte LBP, SIFT özellik çıkarım yöntemine göre daha başarımı yüksek sonuçlar üretebildiği gösterilmiştir [7]. Özellik çıkarımında, hızlı olması ve komşu pikseller arasındaki ilişkinin incelenebilmesinden dolayı Histogram tabanlı özellik çıkarım yöntemi de kullanılmaktadır [8].…”

Section: Gi̇ri̇ş (Introduction)unclassified

“…Gri seviye Histogramı renk düzeylerini içeren toplam nokta sayısını verir. Histogramın uygulama aşamaları, aşağıda adım adım verilmiştir.SIFT detektörü farklı görünümler arasında güvenilir eşleştirme gerçekleştirmek için kullanılabilecek bir anahtar nokta tanımlayıcısıdır[7]. Bu anahtar nokta tanımlayıcısı döngüsel, boyutsal, parlaklık, ölçek ve gürültüye karşı dirençli olduğundan yaygın bir şekilde kullanılmaktadır.…”

unclassified

“…ELM' de giriş ağırlıkları rastgele üretilirken, çıkış ağırlıkları tek bir çevrimde analitik yöntemler ile hesaplanır. Bu yöntem ile eğitim süreci ANN sınıflandırıcısına göre daha hızlı şekilde sınıflandırmayı tamamlamaktadır[7,9].…”

unclassified

See 2 more Smart Citations

Grafiksel Arayüz Tabanlı Mermer Sınıflandırma Uygulaması Geliştirme

Turan¹,

Dandil

Bal

2020

Gazi Üniversitesi Fen Bilimleri Dergisi Part C: Tasarım Ve Teknoloji

Self Cite

View full text Add to dashboard Cite

This study focuses on the development of a graphical interface application that automates the marble classification process by removing it from the manual structure. The texture classification process of marbles is a problem that can be overcome by using image processing and machine learning-based technologies together. In the scope of the study, a graphical interface that classifies marbles automatically has been developed. The training set is created by extracting feature extraction of marble images classified with GUI. With classifiers using the training set, marble images of unknown class are classified without the need for expert staff. The main screenshot of the developed GUI is given in Figure A. Figure A. GUI homepagePurpose: Within the scope of the study, it was aimed to develop a system that automates the marble classification process by removing personnel dependency. Histogram, LBP and SIFT were examined in terms of feature extraction and added to the application. As the classification algorithm, Extreme Learning Machine (ELM), Decision Tree (DT), Support Vector Machine (SVM) and Artificial Neural Network (ANN) were examined and used in the application. Theory and Methods:The two basic units of the marble classification process, feature extraction and classification, are explained step by step. The effect of input parameters used in classification algorithms on classification success is examined and given on the GUI. GUI can be a basis for industrial applications by giving software design architecture. Feature extraction processes are Python based, GUI and classification algorithms are developed based on Java. In the feature extraction process, advanced Python mathematical functions and Swing, Java's advanced interface development library, were used. In addition, by using Weka's Java library for classification functions, classification functions can be integrated into GUI in an easy and modular way.Results: According to the proposed feature extraction and classification algorithms, by learning the images of marble species belonging to 4 different classes, 20 from each, 90-95% crossverified classification results were achieved. After these results are obtained, the GUI 'design architecture has been removed and developed for the end user. Conclusion:The developed GUI offers a solution to a big problem for industrial use. It is aimed to reach the most accurate result by determining the class according to the majority result by running 4 different classifiers simultaneously. If the features of the marble image whose class is determined do not meet the company's criteria, the training set development module developed on the GUI can be automatically added to the training set with its class and features. This enables training sets to be created that provide higher classification performance without any workload.

show abstract

A novel marble recognition system using extreme learning machine with LBP and histogram features

Turan

Uçar

Dandil

2021

Concurrency and Computation

Self Cite

View full text Add to dashboard Cite

Summary Marble classification in production facilities is a sensitive application, which results in light of the subjective decisions of experts. The expert classifies marble manually with its color, homogeneity, and texture in the process. An intelligent marble classifier based on image processing can provide solutions to current problems of the industry. In the proposed study, we introduce an intelligent classifier for marble classification with different classes in real field production. The purpose of the proposed intelligent model for marble facilities is to automate and enhance the manual classification process at present. The real‐world dataset consists of Rosso‐Levanto, Onyx, Keivan, and Black marble images. Local Binary Patterns and Histogram are used for feature extraction and Extreme Learning Machine is designed as an intelligent classifier. Decision Tree, Support Vector Machine, and Artificial Neural Network structures are also used for thorough performance analysis. The findings (successful test rate of 97.5%) reveal a high performance comparing to existing studies.

show abstract

Local Binary Patterns and Extreme Learning Machine based Texture Classification of Marbles

Cited by 4 publications

References 4 publications

CMLBPIncoherent: a New Contextual Image Descriptor for Scene Classification

CMLBPIncoherent: a New Contextual Image Descriptor for Scene Classification

Grafiksel Arayüz Tabanlı Mermer Sınıflandırma Uygulaması Geliştirme

A novel marble recognition system using extreme learning machine with LBP and histogram features

Contact Info

Product

Resources

About