Rotation invariant color texture classification using multiple sub-DLBPs

• DLBP descriptor is presented to characterize the color texture information.
• Create an index image to depict the color information.
• Combine the color index value with DLBP by estimating their joint distribution to construct a MS_DLBP descriptor.
• Discuss about the effect of different color spaces in color texture classification using the proposed MS_DLBP descriptor.
• Experimental results show that the MS_DLBP descriptor is rotation invariant and is more discriminative.

It is known that the rotations of real-world color textures will vary arbitrarily. This paper presents a novel, simple, yet powerful method for rotation-invariant color texture classification. Firstly, we define a Distance-based Local Binary Pattern (DLBP) descriptor to characterize the color texture. By learning the joint distribution of the rotation-invariant DLBP and color intensity information, we define our Multiple Sub-DLBPs   (MS_DLBPMS_DLBP) descriptor. The MS_DLBPMS_DLBP features defined in this paper are invariant to rotation. Here, we also compared seven important color spaces in terms of their effectiveness in our proposed MS_DLBPMS_DLBP approach. The experimental results on the Outex and CUReT databases show the defined DLBP descriptor performs better than the existing color LBP descriptors and the proposed MS_DLBPMS_DLBP approach is very robust to rotation invariance and outperforms state-of-the-art texture analysis methods. Also, HSV color space is shown to outperform the other color spaces in many cases.