Abstract

This chapter discusses the process of object classification with the very useful maximum-likelihood method. One straightforward and quite powerful approach is the use of the Bayes maximum-likelihood classifier. It generates second-order surfaces that partition the color space into disjoint regions, one for each object type (i.e., for each color of pixel). Assuming Gaussian distributions for the clusters of points in color space, the Bayes classifier maximizes the probability that each pixel will be assigned correctly. The first step is to train the classifier to recognize the three types of pixels. This requires a training set containing pixels that are known to fall in the background, inside normal cells, and inside abnormal cells. It is the statistics of these training set pixels that constitute the knowledge the classifier has about the problem. Estimating these statistics is the process of classifier training.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call