Publication

Abstract : Presented herein is an enhanced bi-dimensional empirical mode decomposition (BD-EMD) technique incorporating radial basis function with Hermite interpolation (RBF-HI) for image decomposition. Random point interpolation is an integral aspect of BD-EMD. The characteristics of the two-dimensional intrinsic mode functions (2D-IMFs) derived from BD-EMD significantly hinge on the employed interpolation technique. This study systematically examines various interpolation methods, advocating for the adoption of the RBF-HI technique to augment the performance of BD-EMD. Through comprehensive analysis and comparison of diverse interpolation techniques, crucial features of BD-EMD, including precise surface interpolation and image reconstruction, are refined for enhanced optimization. The outcomes attest that the proposed RBF-HI technique outperforms other methods in terms of surface interpolation, demonstrating superior curve fitting and reconstructed image quality. A standard test image of Lena is utilized for performance benchmarking against existing methods. Comparative analysis concludes that the proposed method excels not only in image decomposition but also in image reconstruction. Performance metrics, encompassing processing time, algorithmic complexity, and peak signal-to-noise ratio (PSNR), underscore the efficacy of the proposed method as a superior approach to BD-EMD for image analysis.