| paulusuni-koblenz. de (2007) | |||||||||||||||
Abstract | |||||||||||||||
| vogtin ormat ik. uni-erlangen. de In this paper we present a new approach for color texture classification which extends the gray level statistical geometrical features (SGF) proposed by Chen. This feature extractor computes 16 statistical measures based on the geometrical properties of connected regions in a series of binary images derived from the gray scale image. We propose intra-plane, inter-plane and the so-called non-linear H'V features as extensions into the color domain. Improvements in classification rates up to 20 % by using our multispectral statistical geometrical features compared with the original gray scale features and color histograms have been achieved. | |||||||||||||||
Publication details | |||||||||||||||
| |||||||||||||||