Image interpolation with self-training using wavelet transform and neural network
Ching-Lin Li, Kuo‐Sheng Cheng · 2008
Interpolation plays an important role in static images and video sequences analysis. High resolution provides important information about still images or video sequences. In this paper, a novel method that combines the wavelet transform and neural network is proposed for image interpolation. Haar wavelet transform and multilayers perceptron are applied. In order to evaluate the image quality, PSNR and a new image quality are both computed for the interpolated images. From the experimental results of testing five images, the proposed method may produce a better image quality of interpolated images than those for the other two traditional methods such as bilinear interpolation and bicubic interpolation.