Texture image analysis based on joint of multi directions GLCM and local ternary patterns
This is an incremental improvement for texture analysis in computer vision, potentially aiding applications like object detection.
The paper tackles texture classification by combining local ternary patterns and gray-level co-occurrence matrix features, achieving higher accuracy on the Brodatz benchmark dataset compared to some state-of-the-art methods.
Human visual brain use three main component such as color, texture and shape to detect or identify environment and objects. Hence, texture analysis has been paid much attention by scientific researchers in last two decades. Texture features can be used in many different applications in commuter vision or machine learning problems. Since now, many different approaches have been proposed to classify textures. Most of them consider the classification accuracy as the main challenge that should be improved. In this article, a new approach is proposed based on combination of two efficient texture descriptor, co-occurrence matrix and local ternary patterns (LTP). First of all, basic local binary pattern and LTP are performed to extract local textural information. Next, a subset of statistical features is extracted from gray-level co-occurrence matrixes. Finally, concatenated features are used to train classifiers. The performance is evaluated on Brodatz benchmark dataset in terms of accuracy. Experimental results show that proposed approach provide higher classification rate in comparison with some state-of-the-art approaches.