Feature Extraction with Color Texture-Sensitive, Rotational, and Scaling Invariant Capability Using Eigenvector-guided Self-organizing Mapping
Invariant scaling and rotational recognition of an image has been successfully realized by extracting the features of the image based on various techniques such as moment, e.g. Zernike moment, pulsed coupled neural network, and high order neural network. These approaches are costly in terms of computational time and network complexity. They are not practical when applied with an image of size at least 256 × 256 pixels. In addition, the problem of extracting the colored texture of an image is also essential in image retrieval application. Although a technique of color metric based on Kubella-Munks photometric reflectance can efficiently capture the color distribution of an image, it does not capture and consider the actual texture of the image. In this paper, we consider both problems and propose an approach which simultaneously reduces the computational and network complexity and captures the colored texture of an image. Our approach is based on the concept of self-organizing mapping network such as Kohonens competitive learning with eigenvector guided feature. Our experiments show that the proposed technique can distinguish colored images with different textures under scaling and rotational aspects while the other techniques cannot.
K. Sookhanaphibarn C. Lursinsap
Advanced Virtual and Intelligent Computing Center Department of Mathematics Chulalongkorn University Bangkok 10330, Thailand
国际会议
8th International Conference on Neural Information Processing(ICONIP 2001)(第八届国际神经信息处理大会)
上海
英文
539-544
2001-11-14（万方平台首次上网日期，不代表论文的发表时间）