Texture image Classification based on improved local Quinary patterns
Texture image classification is an active research topic in computer vision that play an important role in many applications such as visual inspection systems, object tracking, medical image analysis, image segmentation, etc. So far, there are many descriptors for texture image analysis such as local binary patterns (LBP). LBP is a nonparametric operator, which describes the local spatial structure and the local contrast of an image. Local quinary patterns (LQP) is one of the improved versions of LBP in terms of classification accuracy. Statistic input parameters and don’t providing significant binary patterns are some disadvantages of LQP. In this paper a new version of LBP is proposed, which is known as improved local quinary patterns (ILQP). In this paper, a new definition is proposed to divide local quinary codes to four binary patterns. Each extracted binary patterns represent a subset of local features. Also, a new algorithm is proposed here to provide dynamic thresholds in dividing process of LQP. The proposed approach is evaluated using Outex, and Brodatz data sets. Our approach has been compared with some state-of-the-art methods. It is experimentally demonstrated that the proposed approach achieves the highest accuracy in comparison with most of the state-of-the-art texture classification approaches. Low computational complexity, rotation invariant, low impulse-noise sensitivity and high usability are advantages of the proposed texture analysis descriptor.