Our research was supported by grants from the National
Science Foundation, National Institute of Health
and the New Jersey Commission on Science and Technology.
The publications are classified by their main topics and are
listed in all the categories that apply.
A reverse chronological listing of all the available publications
Estimation under Heteroscedasticity
In heteroscedastic data a different covariance matrix is associated with each data point,
a situation which appears in many vision tasks due to the geometric constraints.
Robust Analysis of Visual Data
Techniques to handle complex data having multimodal empirical distributions or
being corrupted by outliers.
Bootstrap as a Tool for Computer Vision
Bootstrap is a resampling method which allows to
infer properties related to the sampling distribution of an estimator from
the available single input.
We are addressing only specific applications (like diagnostic pathology)
and performance related issues.
Everything else: color processing, invariants.