Heuristics for grey-weighted distance computations
2010 (English)In: Symposium on Image Analysis, Uppsala, March 11-12. Proceedings SSBA 2010., 2010Conference paper (Other academic)
With new imaging techniques and computing power come increasing demands on the computation cost of image analysis algorithms. Grey-weighted distance transforms are traditionally computed using graph-based region-growing algorithms. These algorithms expand nodes in all directions by evolving an isotropic wave front. When calculating point-to-point distances, the isotropic propagation can be unnecessarily costly since it expands many nodes in directions away from the goal node. By introducing a heuristic to guide the search, the number of excessive nodes can be decreased. Here we introduce heuristics for computing Distance on Curved Spaces and Weighted Distance on Curved Spaces. We also examine the impact of these heuristics together with a heuristic previously proposed for fuzzy distance computations. The results show that the number of nodes expanded in point-to-point grey-weighted distances can be decreased by up to ~79%.
Place, publisher, year, edition, pages
Computer and Information Science
Research subject Computerized Image Analysis
IdentifiersURN: urn:nbn:se:uu:diva-121572OAI: oai:DiVA.org:uu-121572DiVA: diva2:305768