Modern microscopy techniques allow to acquire temporal sequences of 3D stacks. It provides a unprecedented means to image developing organs or embryos at the cellular level, but results in a huge quantity of data. Proposing efficient processing methods is then absolutely crucial. We are addressing here the study of ascidian embryo development. The stereotypy of their development enables population studies at a cellular level but requires embryo cells to be named. Cell naming can be achieved by name transfer from a database, which, in turn, requires embryo-to-embryo registration, that is eased by embryo symmetry detection. In this article, we propose a novel symmetry detection method for that purpose.