Using global information to refine local patterns for texture representation and classification
作者：
Highlights：
• This paper proposes a novel global refined local binary pattern (GRLBP) by analyzing the nature of the distribution of pixel intensity in local neighborhoods.
• GRLBP consists of two descriptors termed as magnitude refined local sign binary pattern (MRLBP_S) and center refined local magnitude binary pattern (CRLBP_M).
• MRLBP_S distinguishes local neighborhoods with contrast differences by using global magnitude anchors to refine local sign patterns.
• CRLBP_M identifies local neighborhoods with gray differences by employing global central gray anchors to refine local magnitude patterns.
• RLBP has obvious advantages in classification performance, computational complexity, and feature dimension.
摘要
•This paper proposes a novel global refined local binary pattern (GRLBP) by analyzing the nature of the distribution of pixel intensity in local neighborhoods.•GRLBP consists of two descriptors termed as magnitude refined local sign binary pattern (MRLBP_S) and center refined local magnitude binary pattern (CRLBP_M).•MRLBP_S distinguishes local neighborhoods with contrast differences by using global magnitude anchors to refine local sign patterns.•CRLBP_M identifies local neighborhoods with gray differences by employing global central gray anchors to refine local magnitude patterns.•RLBP has obvious advantages in classification performance, computational complexity, and feature dimension.
论文关键词：Texture classification,Texture descriptor,Texture representation,Feature pattern refinement,Local binary pattern
论文评审过程：Received 4 August 2021, Revised 7 June 2022, Accepted 9 June 2022, Available online 15 June 2022, Version of Record 30 June 2022.
论文官网地址：https://doi.org/10.1016/j.patcog.2022.108843
