Skip to content

Angular_predict_continuous

Adrian Quintana edited this page Dec 11, 2017 · 1 revision

angular_continuous_assign

Purpose

This program assigns Euler angles to experimental projections by minimizing the difference between the given experimental image and the central slice of a reference volume in Fourier space. All interpolations are based on a B-spline model of the images and the volume. The translations are also optimized. Since an iterative optimization is performed, it must be initialized with some rough estimate of the angles. The output of xmipp_angular_predict can be used as initialization without any transformation.

The translations or rotations in the header of the images are never applied. Instead, they must be already taken into account in the initialization. This is so, if the output of xmipp_angular_predict is used.

The output of this program is a document file with the predicted angles for each experimental image. The format of this file is


; Predicted_Rot Predicted_Tilt Predicted_Psi Predicted_ShiftX Predicted_ShiftY Corr


This file may be input to xmipp_angular_distance or xmipp_docfile_histogram to compute statistics about the angular assignment step.

Usage


$ angular_continuous_assign -i [selfile in] ... 


Parameters

  • `` Set of files and initial estimates. Note that the initial alignment parameters are taken from this file and not from the image headers
  • `` Reference volume
  • `` Do not produce on screen information
  • `` Document file with the initial estimates of the angles of the experimental projections
  • `` Document file with the predicted angles of the experimental projections
  • -gaussian_Fourier [w0.5] = Weighting sigma in Fourier space. Small values of this parameter concentrate the optimization on low frequencies. This value should be below 1
  • -gaussian_Real [w0.5] = Weighting sigma in Real space. Small values of this parameter concentrate the optimization on the center of the image. This value should be below 1
  • -max_iter [it60] = Maximum number of iteration steps. Convergence might be reached before this number of iterations
  • -max_shift [s-1]= Use this option to limit the maximum shift that the optimization process can find. If the minimum is found beyond this limit (note it is an absolute limit on the shift and it is not relative to the initial shift), then the solution given by the initial guess is kept since the optimized one looks suspicious.
  • -max_angular_change [ang360]= Use this option to limit the maximum angular change with respect to the initial solution that the optimization algorithm can find. If the solution found is beyond this limit, then the initial solution is returned instead of the optimized one since this latter looks suspicious.

Examples and notes

A typical use of this program is


$ angular_continuous_assign -i anglesFromDiscreteAssignment.doc -ref reference.vol -oang assigned_angles.txt


--Main.AlfredoSolano - 29 Jan 2007

Clone this wiki locally