DOE OSTI · 956351
Stack filter classifiers
Abstract
Just as linear models generalize the sample mean and weighted average, weighted order statistic models generalize the sample median and weighted median. This analogy can be continued informally to generalized additive modeels in the case of the mean, and Stack Filters in the case of the median. Both of these model classes have been extensively studied for signal and image processing but it is surprising to find that for pattern classification, their treatment has been significantly one sided. Generalized additive models are now a major tool in pattern classification and many different learning algorithms have been developed to fit model parameters to finite data. However Stack Filters remain largely confined to signal and image processing and learning algorithms for classification are yet to be seen. This paper is a step towards Stack Filter Classifiers and it shows that the approach is interesting from both a theoretical and a practical perspective.
Explore related subjects
Keep this discovery
Explore connections, maps & timelines
Porter, Reid B, Hush, Don. 2009-01-01. Stack filter classifiers. https://doi.org/10.1007/978-3-642-03613-2_26
Cite the original work for its findings. Save a collection to share your selection of sources.