| 摘要 |
A method for detecting geometrical structures in images, especially in images of chemical and/or biological samples, such as images of cells, the method comprising the following steps: detecting a boundary line of the image object; defining at least one sector inside the image, the origin (pk) of which lies on the boundary line; transforming the image section defined by the sector into a transformation space by means of a transformation that associates signatures in the transformation space with respective geometrical structures in the image section; determining the presence of at least one signature inside the transformation space; and retransforming the signatures from the transformation space into the sector for the representation of the geometrical structure.
|
