Colour morphological operators based on formal concept analysis

Lulu Zhao, Junping Wang, Yanbo Li

Abstract: To avoid the destruction of the colour topology of image and the loss of details caused by the existing morphological operators, new colour morphological operators are proposed based on formal concept analysis (FCA). The main idea is to define a set of structuring elements (SEs) for morphological operators, one for each pixel of the image, which can be adapted to the image content. FCA tools, which have good performance in structuring information, are used to find similar pixels that can form SEs. Then, the new operators are defined and applied to different colour spaces. Experimental results show that the proposed operators outperform other kind of operators in preserving detail information and the structures within colour images, which improve the precision of image processing.