作者: B. Rieger , L. J. van Vliet
DOI: 10.1007/978-3-540-45179-2_3
关键词:
摘要: In this paper we present new insights in methods to solve the orientation representation problem arbitrary dimensions. The gradient structure tensor is one of most used descriptors local multi-dimensional images. We will relate its properties double angle method 2D and Knutsson mapping three or higher a general scheme reduce dimensionality mappings needed derive some these reduced mappings.