xmipp3.protocols.protocol_kerdensom module

class xmipp3.protocols.protocol_kerdensom.KendersomBaseClassify(**kwargs)[source]

Bases: ProtClassify2D

Class to create a base template for Kendersom and rotational spectra protocols that share a common structure.

convertInputStep()[source]
createOutputStep()[source]

Store the kenserdom object as result of the protocol.

rewriteClassBlockStep()[source]
class xmipp3.protocols.protocol_kerdensom.XmippProtKerdensom(**args)[source]

Bases: KendersomBaseClassify

Classifies a set of images using Kohonen’s Self-Organizing Feature Maps (SOM) and Fuzzy c-means clustering technique (FCM) .

The kerdenSOM algorithm anneals from an initial high regularization factor to a final lower one, in a user-defined number of steps.

KerdenSOM is an excellent tool for classification, especially when using a large number of data and classes and when the transition between the classes is almost continuous, with no clear separation between them.

The input images must be previously aligned.