Dual-Tree Complex Wavelet Threshold Denoising Method with Edge Enhancement

This study proposed a dual-tree complex wavelet threshold denoising method with edge enhancement to address the poor edge preservation and low contrast of denoising image in existing Gaussian noise removal methods. The proposed method employed the exceptional dual-tree complex wavelet properties such as translation invariance and multidirectional selectivity. Thus, an adaptive threshold denoising model of dual-tree complex wavelet was presented based on the mathematical model of Gaussian noise and assumption. Furthermore, a multi-directional gradient operator was proposed to obtain an edge image from the denoising image using the threshold denoising model. The edge image was linearly parametrically superimposed on the denoising image to achieve the final denoising image with edge enhancement. The experimental results verify that the proposed method performs well in noise reduction and edge preservation. It also demonstrates that the proposed method has a higher computational efficiency.