No abstract

No abstract

Performance of the image denoising algorithms can be evaluated as a function of two main post-filtering parameters-the noise remained (in the image) known as 'residual noise' and the blurring (or other artifacts) introduced in the image during reconstruction known as 'collateral distortion'. This paper proposes a new model for performance evaluation of image denoising algorithms based on the parameters mentioned above. The proposed model evaluates the quality of the filtered image by taking into account the degree of noise cancellation, different distortions that can be introduced during reconstruction, along with the detail preservation. Simulations are performed on filtered images obtained after deteriorations due to impulse, Gaussian and speckle noises respectively. The obtained results using the proposed model correlates well with the different quality evaluation indices and hence efficiently evaluate the quality of the denoised image.