Features and Classification Methods

  • Niels Haering
  • Niels Da Vitoria Lobo
Part of the The International Series in Video Computing book series (VICO, volume 2)


In this chapter, we discuss alternative classification methods, ways to determine the relevance of features, and feature de-correlation. Since the feature extraction stage is the slowest of the modules, we consider classification without preprocessing in the first part of this section. Next we consider linear, quadratic, and eigen-analysis techniques for the determination of good subsets of features and classification, we will argue that the best method to select subsets of features is based on the same non-linear method as that used for classification. There is little benefit to use one approach to decide on the importance of the features and another approach to classify images. Some systematic approaches are presented which show that the best solutions involve the consideration of an exponential number of combinations of features. The results of the various classification methods are shown in Chapter 4.


Feature Extraction Method Convolutional Neural Network Feature Extraction Stage Quadratic Classifier Steerable Filter 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Springer Science+Business Media Dordrecht 2001

Authors and Affiliations

  • Niels Haering
    • 1
  • Niels Da Vitoria Lobo
    • 2
  1. 1.DiamondBack Vision, IncRestonUSA
  2. 2.School of Electrical Engineering and Computer ScienceUniversity of Central FloridaOrlandoUSA

Personalised recommendations