Selection of key features for invariant object recognition using fuzzy entropy
Xiaofan Liu, Shaohua Tan, V. Srinivasan, Sim Heng Ong · 2002
The paper proposes a novel method of feature selection that is supported by an adaptive segmentation technique using annular and sector windows. To achieve the feature dimension reduction, an adaptive method of selecting key features based on a fuzzy entropy is introduced. The method is applied to a map recognition task to demonstrate its efficacy.>