On Estimation of the Number of Image Principal Colors and Color Reduction through Self-Organized Neural Networks

摘要: A new technique suitable for reduction of the number colors in a color image is presented this article. It based on use Principal Color Components (PCC), which consist components and additional ex- tracted with proper spatial features. The features are used to enhance quality final image. First, principal each PCC extracted. Three algorithms were developed tested pur- pose. Using Kohonen self-organizing feature maps (SOFM) as clas- sifiers, obtained look-up table, containing PCC, con- structed. extracted from table entries through SOFM by setting output neurons equal original To speed up entire algorithm reduce memory requirements, fractal scanning subsampling employed. method independent scheme; it applicable any type images can be easily modified accommodate Several experimental comparative results exhib- iting performance proposed presented.