Image Denoising Based on Curvelet Transform and Continuous Threshold

Curvelet transform is more suitable than wavelet transform for planar image processing. The theory of curvelet transform is introduced. Noise-image is carried on decomposition based on curvelet transform, and distribution characteristics of noise are analyzed. Applying a quantization method of using threshold of which the function is continuous and differentiable is proposed, to remedy disadvantages of quantization methods of using traditional thresholds. Then the method of image denoising is confirmed. The experimental results show that applying the proposed approach can obtain better quality, compared with other methods.