An efficient attribute-space connected filter on graphs to reconstruct paths in point-clouds

Highlights：

• We believe that our method is a starting point to develop a new theoretical base for image analysis and understanding, where we investigate the common ground between the graph- and image-processing techniques. On the other hand, we explore a rather uncommon yet intriguing application area.

• These three papers in the published literature set a theoretical base for application of morphology to either graphs or images. [1] M.H.F. Wilkinson, Attribute-space connectivity and connected filters. Image Vis. Comput. 25 (2007) 426–435. [2] A. Toet, H.J.A.M. Heijmans, P. Nacken, L. Vincent, Graph morphology. J. Vis. Commun. Image Represent. 3 (March 1992) 24–38. [3] Jean Cousty, Laurent Najman, Fabio Dias, Jean Serra. Morphological filtering on graphs. Comput. Vis. Image Understand. 117(4) (2013) 370–385. Special issue on Discrete Geometry for Computer Imagery. In our work, we use the combination of both concepts to introduce a theoretical extension and a very efficient method to process high dimensional and complex images.

摘要

•We introduce a new representation and processing scheme for computing attribute-spaces and attribute-space connected filters. We have extended this concept to process graphs; this is a powerful tool when it comes to analysing images generated by multi-sensor systems. In addition, our method requires intrinsically a smaller amount of memory to store image representations in different attribute-spaces. The latter makes it possible to process more complex and larger images in higher dimensions.•We believe that our method is a starting point to develop a new theoretical base for image analysis and understanding, where we investigate the common ground between the graph- and image-processing techniques. On the other hand, we explore a rather uncommon yet intriguing application area.•These three papers in the published literature set a theoretical base for application of morphology to either graphs or images. [1] M.H.F. Wilkinson, Attribute-space connectivity and connected filters. Image Vis. Comput. 25 (2007) 426–435. [2] A. Toet, H.J.A.M. Heijmans, P. Nacken, L. Vincent, Graph morphology. J. Vis. Commun. Image Represent. 3 (March 1992) 24–38. [3] Jean Cousty, Laurent Najman, Fabio Dias, Jean Serra. Morphological filtering on graphs. Comput. Vis. Image Understand. 117(4) (2013) 370–385. Special issue on Discrete Geometry for Computer Imagery. In our work, we use the combination of both concepts to introduce a theoretical extension and a very efficient method to process high dimensional and complex images.