Texture feature extraction based on histogram of oriented gradient domain texture tendency for tyre pattern retrieval
Yuxiang Ge, Ying Liu, Fuping Wang, Yi Lai, Tingge Zhu, Dianwei Wang · 2017
A novel local texture feature extraction algorithm is proposed based on histogram of oriented gradient domain texture tendency (HOGTT). Classical HOG descriptor proves to be sensitive to rotation transformation, though it is insensitive to illumination and scale changes. The intrinsic texture tendency of tyre tread image has obvious consistency and is robust to different types of images transformation. This paper mainly makes two contributions: First, the scale and circular region constraint on tyre pattern in local feature extraction highly improves the representation precision of the extracted HOG feature. Second, the HOG feature matrix is modified by texture tendency to obtain the rotation invariance property. Experimental results on tyre tread dataset demonstrate the effectiveness of HOGTT in tyre pattern retrieval.