Image Classification Using Synchronized Rotation Local Ternary Pattern

Abstract

In this paper, a synchronized rotation local ternary pattern (SRLTP) operator for image classification is presented. The proposed SRLTP descriptor improves on the local ternary pattern (LTP) method with an additional process on the generated lower and upper LTPs. The lower and upper patterns are encoded to a rotation invariant pattern histogram and a uniform pattern histogram, respectively. Therefore, the feature vector can utilize the advantages offered by the rotation invariant pattern histogram while retaining the original information in the uniform pattern histogram. In addition, a two-dimensional discrete wavelet transform (DWT) and a discrete Fourier transform (DFT) enhanced the robustness of the texture classification are applied in the experiments. The DWT adapts the context variations and reduces noise by applying a high-pass filter and a low-pass filter with different image resolutions. The DFT is mostly used to overcome the rotation variation problem. To verify the performance of proposed SRLTP descriptor, the effectiveness and robustness of the proposed descriptor were compared with those of existing descriptors by using four public texture data sets. The experimental results demonstrate that the performance of the SRLTP descriptor is convincing and outperformed to the existing descriptors.