Evita / News: Recent posts

After a 16-year hiaitus, Evita is back! The code has been patched to compile against modern versions of VTK, QccPack, and FLTK, and a number of bug fixes have been made to get the Evita system operational once more.

With this release, feature detection is now denoised using a multiscale
feature detector. That is, features are detected within basebands at various
resolutions, and those that persist across all scales are retained.
Additionally, in the 2D swirl feature detector, Sperner's Lemma is now
employed to reduce the number of features identified. Specifically, those
features identified as swirls based on the swirl parameter are discarded
unless they also contain a vortex core.... read more

Evita now fully supports the coding and visualization of datasets with both
rectangular and curvilinear grids. Rectangular grids are implicit; that is,
an identifier in the ds file header indicates the presence of a rectangular
grid, but this grid is not coded in the ds file. Curvilinear grids, on the
other, are explicitly included in the ds file. During compression, the
displacement, or "difference," between the curvilinear grid and an implicit
uniform grid is encoded. During decoding and reconstruction, the current grid
difference is added to the uniform grid to generate the current
reconstruction of the grid. The grid local to an individual ROI or ROI tile
is coded and reconstructed independently of others.... read more

At long last, here is the new version of Evita. Substantial changes have
been made since the last release:
- Large ROIs can be split into multiple tiles during encoding; the user
selects the tile size as part of the encoder GUI.
- 2D visualization in the client no longer uses vtkImageActor; rather
vtkStructuredGrid is now used, which will eventually allow
non-rectilinear grids (coming soon).
- Each tile of each ROI is visualized with its own, independent actor
- A transparency mask is included at the start of the encoding for
each tile of each ROI; this means the segmentation map is no
longer communicated between transcoder and client.
- evInteractorStyleImaging has been added; 2D client visualization now
supports rubber-band selection of view area.
- Various bugs have been fixed.... read more

Here is the the next version of Evita.

Essentially, the changes in this release are minor changes for increased build portability. Additionally, VTK version 4.0 is now the supported VTK release.

Evita now has initial support for feature detection. The program dsdetect can now detect swirls, expansion fans, and shocks in CFD datasets. Classes added to support this capability include evFeatureDetector, evFeatureDetectorSwirl, evFeatureDetectorExpansionFan, and evFeatureDetectorShock.

This version of EVITA is the one that was demoed at the EVITA Technical Review and Symposium in June, and is perhaps the first version to include a mostly complete implementation of the server and client.

An EVITA miniconference has been schedule June 20 at MSU ERC. This one day meeting will feature presentations by a multitude of EVITA team members on current EVITA research.