Submitted to CVPR ' 99 Bayesian based Optimal Nearest Neighbor
نویسنده
چکیده
Nearest neighbor rules are popular classiiers, partly due to their good asymptotic properties. Improving the performance of NN rules when only nite samples are available has been studied over the last two decades. For example, smart distance measures which are data dependent have been proposed. From a theoretical point of view, one is interested in deriving the optimal NN distance measure. Existing optimal measures are obtained by minimizing the diierence between-nite sample based NN risk/error and the asymptotic NN risk which is less than twice the Bayes risk. In this paper, we propose new optimal measures based on posterior probability. Using the new distance measures, NN rule can achieve the Bayes risk asymptotically. In addition, Bayesian based robust distance measures are proposed as practically good measures.
منابع مشابه
The Optimal Distance Measure for Object Detection
We develop a multi-class object detection framework whose core component is a nearest neighbor search over object part classes. The performance of the overall system is critically dependent on the distance measure used in the nearest neighbor search. A distance measure that minimizes the mis-classification risk for the 1-nearest neighbor search can be shown to be the probability that a pair of ...
متن کاملAdaptive Metric nearest Neighbor Classification
Nearest neighbor classification assumes locally constant class conditional probabilities. This assumption becomes invalid in high dimensions with finite samples due to the curse of dimensionality. Severe bias can be introduced under these conditions when using the nearest neighbor rule. We propose a locally adaptive nearest neighbor classification method to try to minimize bias. We use a Chisqu...
متن کاملNearest Neighbor Search for Relevance Feedback
We introduce the problem of repetitive nearest neighbor search in relevance feedback and propose an efficient search scheme for high dimensional feature spaces. Relevance feedback learning is a popular scheme used in content based image and video retrieval to support high-level concept queries. This paper addresses those scenarios in which a similarity or distance matrix is updated during each ...
متن کاملFace Recognition Based on Nearest Linear Combinations
This paper proposes a novel pattern classiication approach, called the nearest linear combination (NLC) approach, for eigenface based face recognition. Assume that multiple prototypical vectors are available per class, each vector being a point in an eigenface space. A linear combination of prototypical vectors belonging to a face class is used to deene a measure of distance from the query vect...
متن کاملA Bayesian Segmentation Framework for Textured Visual Images
This paper presents a new framework for segmentation of textured visual imagery. The proposed method consists of a Bayesian formulation for labeling similar regions. Similarity is defined via texture features obtained by Gabor Wavelets. Multivariate Gaussian distributions are employed to model the feature class-conditional densities, while the Markov process is used to characterize the distribu...
متن کامل