Image compression by nonlinear principal component analysis

In recent years, many methods for image compression have proposed, especially JPEG and MPEG have achieved high compression ratio, but these methods can not restore images completely. In these methods image data are reduced in spatial frequency domain according to human eye property. In this study, we have developed a new method to reduce image data especially in noises of image using a neural network. An advantage of this method is to preserve the quality of image by reducing the noise which is independent of original image data.