Denoising Images by Dual–Tree Complex Wavelet Transform Combined With Meta Heuristic Optimization Algorithms

Abstract

Denoising is a prime objective technique for processing images. Image denoising techniques removes the noises present in an image without interrupting its features and contents. The image gets interrupted by channel or processing noise depending on the applications. Thus, the contaminated noises produce degradable image qualities with respect to subjective and objective approach. To overcome this, image denoising approaches were suggested. In the present research, Dual–Tree Complex Wavelet transform (DTCWT) is utilized to achieve image denoising since they perform multi resolution decomposition by two DWT trees. Soft and hard thresholding methods are used to threshold wavelet coefficients. The present research proposes a novel technique to denoise images which gives image information clearly by thresholding and optimization technique. The optimization is carried through different Meta-heuristic optimization Algorithms Genetic Algorithm (GA) and Grey-wolf optimization (GWO) algorithm. Optimization of threshold value is performed after Bayesian method and the observed output produces better results when compared to other techniques involving Visu shrink, Sure shrink and Bayes shrinkbased on peak signal to noise ratio (PSNR) and visual qualities.

Peningkatan Akurasi Algoritma Support Vector Machine Menggunakan Dual-Tree Complex Wavelet Transform Pada Klasifikasi Citra Katarak dan Normal

“…Transformasi wavelet dua dimensi (2D) diprediksi menjadi versi lanjutan dari transformasi wavelet satu dimensi (1D). Untuk mencapai ini, awalnya transformasi satu dimensi diterapkan disepanjang setiap baris dan kolom pada setiap tingkat dekomposisi struktural (Lavanya et al, 2020). Gambar 4.…”