Flowy — A motion field viewer ############################# Description ----------- Flowy provides a motion field viewer for 2D spatial representations. .. image:: flowy/flowy-1.png :scale: 50% Dowload ------- You can browse the releases_ web directory to get the latest version. .. _releases: /data/flowy Dependencies ------------ The `GTK+`_ library, in its third major version, as well as its associated `Python bindings `__ are required to run the software. .. _GTK+: http://www.gtk.org Data formats ------------ Compatible motion arrays are stored as HDF5 datasets containing simple datasets of integer elements. Several formats are supported, each allows an optional temporal domain, which must be in the first dimension if present. Optical flow ~~~~~~~~~~~~ Shape ``[time ×] 2 × height × width`` In the optical flow format, compatible datasets must have three or four dimensions. The last three dimension respectively represent the motion vector, the height and the width of the optical flow. The first dimension should be equal to two, for the X and Y components of the motion vectors. Motion field ~~~~~~~~~~~~ Shape ``[time ×] vdim × height × width`` In the motion field format, the third to last dimension can be equal to ``(N*2+1)²``, where ``N`` is an integer, to represent grid velocities. Otherwise a specific velocity space must be given as command line parameter, or be present in the dataset attributes. Direction field ~~~~~~~~~~~~~~~ Shape ``[time ×] ndirs × height × width`` In the direction field format, ``ndirs`` directions are assumed to regularly span the [0°,360°) range, unless a dataset attribute is present in the HDF5 file.