Kind Of Local Information Research Articles

Global plus local: A complete framework for feature extraction and recognition

Linear discriminant analysis (LDA) is one of the most popular supervised feature extraction techniques used in machine learning and pattern classification. However, LDA only captures global geometrical structure information of the data and ignores the geometrical structure information of local data points. Though many articles have been published to address this issue, most of them are incomplete in the sense that only part of the local information is used. We show here that there are total three kinds of local information, namely, local similarity information, local intra-class pattern variation, and local inter-class pattern variation. We first propose a new method called enhanced within-class LDA (EWLDA) algorithm to incorporate the local similarity information, and then propose a complete framework called complete global–local LDA (CGLDA) algorithm to incorporate all these three kinds of local information. Experimental results on two image databases demonstrate the effectiveness of our algorithms.

Data‐dependent weighted median filtering based on fuzzy rules

When a signal containing rapid changes, such as an image signal, is degraded by additive noise, a weighted median filter, which is a generalization of the median filter, is useful in its restoration. The weight of the weighted median filter is designed by using training signals, and is designed to be optimal for all training signals. The designed filter, however, cannot be optimal for local training signals. From that viewpoint, this paper presents a data-dependent weighted median filter in which the weight is made variable according to local information from the signal to be processed. A data-dependent weighted filter is of course desirable for highly accurate restoration of signals with mixed properties, as in image signals. In order to realize a high performance data-dependent filter, it is crucial to know accurately the local properties of the signal and to reflect them in the weights determined. In this study, three kinds of local information are used to represent the local properties of a signal, and a method for the derivation of weights using fuzzy inference is established. A data-dependent weighted median filter is constructed employing the proposed method. Its performance and effectiveness are examined in several application examples. © 1998 Scripta Technica. Electron Comm Jpn Pt 3, 81(4): 21–32, 1998