Application of texture features and machine learning methods to grains segmentation in rock material images
The segmentation of rock grains on images depicting bulk rock materials is considered. The rocks’ material images are transformed by selected texture operators, to obtain a set of features describing them. The first order features, second-order features, run-length matrix, grey tone difference matrix, and Laws’ energies are used for this purpose. The features are classified using k-nearest neighbours, support vector machines, and artificial neural networks classifiers. The results show that the border of rocks grains can be determined with above 75% accuracy. The multi-texture approach was also investigated, leading to an increase in accuracy to over 79% for the early-fusion of features. Attempts were made to reduce feature space dimensionality by manually picking features as well as by the use of principal component analysis. The outcomes showed a significant decrease in accuracy. The obtained results have been visually compared with the ground truth. The compliance observed can be considered to be satisfactory.
Paper
Full text
Application of texture features and machine learning methods to grains segmentation in rock material images
Semantic Scholar · Computer Science · 2020
Abstract
The segmentation of rock grains on images depicting bulk rock materials is considered. The rocks’ material images are transformed by selected texture operators, to obtain a set of features describing them. The first order features, second-order features, run-length matrix, grey tone difference matrix, and Laws’ energies are used for this purpose. The features are classified using k-nearest neighbours, support vector machines, and artificial neural networks classifiers. The results show that the border of rocks grains can be determined with above 75% accuracy. The multi-texture approach was also investigated, leading to an increase in accuracy to over 79% for the early-fusion of features. Attempts were made to reduce feature space dimensionality by manually picking features as well as by the use of principal component analysis. The outcomes showed a significant decrease in accuracy. The obtained results have been visually compared with the ground truth. The compliance observed can be considered to be satisfactory.