An Efficient FPGA Implementation of Optimized Anisotropic Diffusion Filtering of Images

Table 2

Quality measures (performance metrics): SSIM (structural similarity) and PSNR (peak signal-to-noise ratio) for the experiments. For each column, the best value has been highlighted for three different noise standard deviations. Our proposed technique OAD (optimized anisotropic diffusion) shows better result except for the SSIM parameter for standard deviation, 20. Comparison has been made with different types of benchmark edge preserving denoising filters.

Method	Std. dev. = 12		Std. dev. = 15		Std. dev. = 20
	SSIM	PSNR (dB)	SSIM	PSNR (dB)	SSIM	PSNR (dB)
(a) ADF [2]	0.9128	29	0.8729	27.82	0.8551	24.93
(b) NLM [27]	0.9346	28.2	0.9067	27.29	0.8732	25.64
(c) BF [12]	0.9277	27	0.8983	28.54	0.8809	24.28
(d) TF [13]	0.8139	25.22	0.7289	22.87	0.6990	21.98
(e) OAD; (Our proposed optimized anisotropic diffusion filter)	0.9424	30.01	0.9245	28.87	0.8621	25.86