Hi Felix,

Might be of your interest. 

http://deepblue.lib.umich.edu/bitstream/handle/2027.42/83684/JCK-JLP2009.pdf?sequence=1

A nice paper and thanks Karthick. But they did multivariate analysis of shapes using PCA manually, which is possible, but too lengthy if you have many samples. May be any simpler GUI software alternatives? 

Have you tried with ImageJ?

http://imagej.nih.gov/ij/

Dear Felix: some colleagues and I have recently developed RPS: Resistant Procrustes Software, an open source (and free) software to perform basic shape analysis on 2D and 3D datasets.

This tool enables you to perform a Procrustes superposition of the shapes you want to study (you have two choices: a least squares superposition criterion, and a resistant superposition criterion; is useful to perform both and compare), to compute the corresponding distances between shapes afterwards, and to obtain an ordination of shapes based on the computed distance matrix (ie, a graph in 2D or 3D where each shape is represented by a single dot, and where the distance you see between dots approximately represents the corresponding shape distance you computed)

Hope this helps. 

Thanks a lot! It seems both Resistant Procrustes Software, as well as Bookstein's Morphometrics will be of great help with my kind of works. I will check them out. And Karthick, ImageJ doesnt do shape analysis I guess (I have used it)

Our Checkpoint software (below) can help you collect landmark points as well as perform some 3D shape analysis on the collected points, including PCA and Procrustes.

https://www.stratovan.com/products/checkpoint?utm_source=rg&utm_medium=ans&utm_campaign=stat1

The geomorph package of R is good for landmark based analysis, and you can create the script in R, then create a shell script to call up, said script with parameter modifications. Also, there is morphoJ if you want a GUI. For 3D statistical shape modeling you can also use Scalismo in Scala or the R package RvtkStatismo.