Please use this identifier to cite or link to this item:
|標題:||Object DetectionUsing Color Entropiesand a Fuzzy Classifier||Project:||Computational Intelligence Magazine, IEEE, Volume8 , Issue 1.||摘要:||
This paper proposes a novel approach tospecific object detection in complex scenes usingcolor-based entropy features and a fuzzy classifier(FC). Appearances of the detected objects are assumedto contain multiple colors in non-homogeneousdistributions that make it difficult to detect theseobjects using shape features. The proposed detectionapproach consists of two filtering phases with twodifferent novel color-based entropy features. Thefirst phase filters a test pattern with the entropy ofcolor component (ECC). A self-splitting clustering(SSC) algorithm is proposed to automatically generateclusters in the hue and saturation (HS) colorspace according to the composing pixels of anobject. The ECC value is computed from histogramsof pixels in the found clusters and is used togenerate object candidates. The second filteringphase uses the entropies of geometric color distributions(EGCD) to filter the object candidatesobtained from the first phase. An EGCD is computedfor each of the clustered composing colors ofa candidate object. The EGCD values are fed to anFC to enable advanced filtering. A new FC usingthe SSC algorithm and support vector machine(FC-SSCSVM) for antecedent and consequentparameter learning, respectively, is proposed toimprove detection performance. Experimentalresults on the detection of different objects andcomparisons with various detection approachesand classifiers verify the advantage of the proposeddetection approach using the FC-SSCSVM.
|Appears in Collections:||電機工程學系所|
Show full item record
TAIR Related Article
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.