Difference between a PCA based model and Reduced model

[Venkatesh Kaduru]

Hello,

I would like to know the conceptual difference between the statismo build shape model and the statismo reduced model. Once we execute the build shape model, we get the PCA components. What happens when we reduce this PCA model, as I have read both the tutorials, reduced and build shape model,  but I am interested to know more apart from the number of components that we can use while building a reduced model. It would be appreciated if someone could explain in detail. 

Note: [ I have a PCA based model with ~ 4k indexes/components. The majority of the shape variations can be seen in the first 5 components with the variance (example, gender, thin, thick, smallest, highest). As per my understanding when increases the indexes the values will be decreased. ]

Thanks,

Venkatesh Kaduru

[Marcel Luethi]

Hi Venkatesh,

Principal component analysis gives you a set of basis vectors, called principal components, which span the space of possible shapes. The basis vectors are ordered by decreasing variance. The reduced model simply discards those principal components which contribute least to the total shape variation spanned by the model. You end up with another PCA model, but one which does not represent all the variation you observed in the training data.

Best regards,

Marcel 

--
You received this message because you are subscribed to the Google Groups "statismo" group.
To unsubscribe from this group and stop receiving emails from it, send an email to statismo-user...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Venkatesh Kaduru]

Hello Marcel,

Sorry for the late response, thanks for clarifying me and also I did see few tutorials on PCA computation on data. 

Thanks,

Venkatesh Kaduru  

On Monday, March 4, 2019 at 8:52:00 AM UTC+1, Marcel Luethi wrote:

Hi Venkatesh,

Principal component analysis gives you a set of basis vectors, called principal components, which span the space of possible shapes. The basis vectors are ordered by decreasing variance. The reduced model simply discards those principal components which contribute least to the total shape variation spanned by the model. You end up with another PCA model, but one which does not represent all the variation you observed in the training data.

Best regards,

Marcel