Savitzky Golay and KPCA based Optimal Discrete Wavelet Transform Architecture for Image Compression

Abstract

• In pre-processing, Savitzky Golay filtering technique (SGFT) is utilized to eliminate the noise signals thereby generating accurate results. • Utilizing DWT along with Kernel principal component analysis and Iterative artificial ecosystem optimization algorithm for post-processing thus obtaining an optimal reconstructed image with high quality. • Comparing our proposed approach with various other techniques to determine the effectiveness of the system. In recent years, the DWT acts as an effective tool in image compression and decomposition. There exist numerous transform techniques for compressing images but there occurs information loss and low picture quality. To overcome such drawbacks this paper aims in developing an optimal reconstructed image with high quality. The proposed approach comprises two significant phases namely the pre-processing and post-processing. During pre-processing, the multispectral images are enhanced by removing the noise signals from the input image. In this paper, a Savitzky Golay filter is employed to eliminate noise signals and smoothing the multi-spectral image for the generation of more accurate results. The pre-processed image is then subjected to post-processing. In post-processing, the DWT along with Kernel principal component analysis and Iterative artificial ecosystem optimization algorithm is employed to obtain an optimal reconstructed image with high quality and without losing the information. In addition to this, the proposed model is trained using aerial images obtained from the SIPI image database. The simulation metrics are evaluated for various approaches and the results reveal that the proposed approach outperforms other approaches.