|
Status
More frequent updates can be found at the Status
Report page.
MileStone: |
GUI implementation |
Time Frame: |
1/04 - 2/04 |
Description: |
The first two versions of CrySis were proof-of-concept implementations.
They accomplished their mission of proving that neural networks could
offer very reasonable accuracy for image analysis in a timely fashion.
The paper
written from that research summarizes my findings. Version 3 of CrySis
will now bring the test version up to the level of a releasable application.
It will feature a fully functional GUI which divides CrySis's current
features into three categories (organized by tabs): image inspection,
diffraction analysis, and for the purposes of optimizing the analysis
algorithms, training. The code has been completely restructured and
modularized. This radical alteration will enable new algorithms and
new image formats to be added with virtually no additional effort.
Thus more time can be spent perfecting the essential features rather
than rewriting basic code. This reconstruction will give CrySis a
new face and enable it to expand into an essential component of high
throughput crystallography. |
rev. 1/20/04
|