Drosophila Collaboration
Presentation
Overview (5 min)
- CCB and Collaborations
- Dr. David Houle, FSU
Project Description (5 min.)
- project description (insert description of Dr. Houle's project here)
- his challenges:
- measure morphologic phenotype efficiently
- visualize changes in phenotype
- study of the relationship between genotype and phenotype
- Our Challenges
- various input data file formats
- minimize special purpose code
- enable him to run his own analyses on his local (Windows and Unix) machines.
- Understanding and Communicating about the problem
- Our answers
- translate various input formats into a common 'landmark' file format
- split functions of (fly) image registration, warp, and warp analysis into separate tools.
- use PERL to coordinate analysis programs (for portability, efficiency analyzing 20,000 files, and per Dr. Houle's request).
- In person meeting with Dr. Houle for two days.
Tools used
- See: Landmark Warp Library
- Types of transformations (2D and 3D unless stated otherwise):
- Rigid
- Procrustes method (3D only)
- Dual Quaternion method (3D only)
- Rigid with Scaling
- Generalized Procrustes method (3D only)
- Affine
- linear least squares best fit
- Non-linear spline
- elastic pde w/ linear or hyperbolic forces (non-radial basis)
- biharmonic equation (thin plate spline radial basis)
- inverse multiquadratic radial basis
- gaussian radial basis
- wendland radial basis
- Example of use (insert sample code here)
- Describe applications created
Results obtained
Booklet entry for fly wing
Presenter
Dr. Roger Woods
Description
Landmark Registration tools, as applied to Drosophila wing image warping
To support a collaboration with Dr. David Houle, we developed a Java library
to simplify creation of landmark warping applications.
This was used in collaboration with Dr. Houle to create tools that compare
the phenotypes of various species of Drosophila. A warp field is derived from homologous fiducial landmarks
in fly wing images. Descriptive statistics of wing area change are computed across each species of fly from
these fields. These tools are not limited to image data, but may be applied to any two or three dimensional arrays of scalars.
The transformations estimated by this library are useful for image registration problems and shape analysis. Supported
transformations include Procrustes, Generalized Procrustes, non-linear spline and affine transformations.
Image
Shape Tools Update
Presentation
Overview (5 min)
- Mission
- Approach
- Portable Code
- Available Code
- Documented
Recent Accomplishments (5 Min)
Libraries (5 Min)
Demonstration (5 Min)
- ShapeViewer -
- Medial Axis pipeline
- Geodesic Distance Pipeline
Announce CD / Questions
Booklet entry for ShapeTools
Presenter
Craig Schwartz
Title
An update of the ShapeTools libraries and applications
Description
The ShapeTools group creates portable scientific software focussed on computational geometry data structures: things composed of points, edges and faces.
We will describe the additions to our supported file formats and applications created since the last CCB meeting, with emphasis on tools of general use to the Neuroimaging and scientific community.
A CD of the current set of ShapeTools libraries and applications will be available.
Image
Image Caption:
The ShapeViewer displaying a cortical surface (grey) and the geodesic path
connecting two vertices of that surface. This path is computed with the
ShapeTools application GeodesicPath, using Dijkstra's algorithm.
Other
FlyWingImage.jpg 
