Xmipp

xmipp_classify_kerdensom (v3.0)

Usage

Purpose
Kernel Density Estimator Self-Organizing Map KerDenSOM stands for Kernel Probability Density Estimator Self-Organizing Map. It maps a set of high dimensional input vectors into a two-dimensional grid. For more information, please see the following reference.

The topology of the network can be hexagonal or rectangular (see below). It is advised to design maps with one of its sides larger than the other (e.g. 10x5).

   Xdim is ------>
   HEXAGONAL:
 O O O O O O O O O
O O O & & & O O O
 O O & @ @ & O O O
O O & @ + @ & O O
 O O & @ @ & O O O
O O O & & & O O O
 O O O O O O O O O
   RECTANGULAR:
O O O O O O O O O
0 O O O & O O O O
O O O & @ & O O O
O O & @ + @ & O O
O O O & @ & O O O
O O O O & O O O O
O O O O O O O O O

See also
image_vectorize

Parameters

-i <file_in>
Input data file This file is generated by image_vectorize
--oroot <rootname>
rootname_classes.xmd, rootname_images.xmd and rootname_vectors.xmd will be created This file mst be read by image_vectorize
--xdim <Hdimension=10>
Horizontal size of the map
--ydim <Vdimension=5>
Vertical size of the map
--topology <topology=RECT>
Lattice topology The following picture will help in understanding the differences between both topologies and the map axis convention:
where <topology> can be:
  • RECT
  • HEXA
--deterministic_annealing <steps=10> <Initial_reg=1000> <Final_reg=100>
Deterministic annealing controls the smoothness regularization Set it to 0 0 0 for Kernel C-means
--eps <epsilon=1e-7>
Stopping criteria
--iter <N=200>
Number of iterations
--norm
Normalize input data

Examples and notes

xmipp_image_vectorize -i images.stk -o vectors.xmd
xmipp_classify_kerdensom -i vectors.xmd -o kerdensom.xmd

User's comments

%COMMENT{type="tableappend"}%