Octave Forge level-set

42 lines (38 with data), 2.2 kB

Version 0.3.1 (trunk):
  * Updated the C++ code to not depend on functions deprecated in Octave 4.2.
  * Modernised the code in general to work well with Octave 5.1.
  * Specified dependency on "parallel" as well as bumped minimum required
    Octave version.

Version 0.3.0 (2015-12-17):
  * Changed the build system to use autoconf.
  * ls_genbasic can now be called with arbitrary coordinates, they need
    not correspond to the full grid.  In particular, single points
    can be passed as arguments as well.
  * Add new routines for shape optimisation with level sets.  The main
    one is so_run_descent.
  * Allow saving and replaying shape optimisation descents.  This is
    accomplished with so_save_descent and so_replay_descent.  Also
    so_explore_descent can be used to interactively explore a descent log.

Version 0.2.0 (2015-02-18):
  * The .m file functions don't check nargout() any longer.  This is checked
    by the interpreter anyway.  Also, usage() as replaced by either
    print_usage() or error().
  * New function 'ls_normalise' that does the normalisation of exactly zero
    values that was previously done only in ls_find_geometry.
  * 'ls_inside' as well as 'ls_equal' (and all other set functions that use
    them internally) now consider negative zeros to be part of the level-set
    domain.  This is consistent with the behaviour of 'ls_normalise'.
  * 'ls_animate_evolution' only clears the contour between frames now,
    so that the "jitter" is drastically reduced.
  * The usage of 'ls_sign_colourmap' was slightly changed.  It allows
    now to specify the colours making up the gradients.
  * New feature: Time stepping of the level-set equation as alternative
    method for evolution.  New methods 'upwind_gradient_norm' and
    'ls_time_step'.
  * New function 'ls_distance_fcn'.  It calculates the non-signed distance
    function, and is faster than 'ls_signed_distance'.  Only exterior points
    are processed with Fast Marching.
  * New function 'ls_hausdorff_dist' to approximate the Hausdorff
    distance between two sets from their distance functions.

Version 0.1.0 (2014-06-09):
  * Initial release.