Numerical Calculation of Partial Differential Equation Deduction in Adaptive Total Variation Image Denoising

2019 
Total variation model of image denoising is easy to influence the gradient and lose the details of image. Due to these weaknesses, many adaptive total variation (ATV) models of image denoising have been proposed to eliminate the Gaussian noisy additive in the image. This paper implements the model through a numerical solution, where the gradient descent method is used to derive the partial differential equation (PDE) corresponding to the ATV model. First, based on Euler-Lagrange equation, a detailed derivation process for Partial Differential Equation is used to calculate the ATV model. Then, based on gradient descent method, a numerical calculation of the model is derived from the PDE using the Direct Difference Method. Finally, several different λ parameters are compared to produce different image denoising effects and the appropriate parameter λ value is determined. Experimental results prove that our proposed numerical calculation can effectively realize the ATV denoising model.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    20
    References
    1
    Citations
    NaN
    KQI
    []