Table 5 A quantitative comparison of the performance of the spiking noise removal platform based on CORDIC_HR model in comparison with other noise removal methods.
Noise removal method | Poisson noise with \(\lambda =1\) | Salt and pepper noise with 10% noise probability | ||
|---|---|---|---|---|
MSE (mean squared error) | PSNR (peak signal-to-noise ratio (in dB)) | MSE | PSNR | |
Arithmetic filter40 | 114.31 | 27.54 | 27.04 | 33.81 |
Geometric filter40 | 24.46 | 34.24 | 30.54 | 33.28 |
Harmonic filter40 | 31.15 | 33.19 | 34.03 | 32.81 |
Contra-harmonic filter40 | 252.15 | 24.11 | 252.12 | 24.11 |
Median filter40 | 19.79 | 35.16 | 9.29 | 38.44 |
Max and min filter40 | 79.63 | 29.11 | 94.15 | 28.39 |
Mid-point filter40 | 16.24 | 36.02 | 15.95 | 36.10 |
Spiking noise removal network35 | 20.8 | 33.40 | 9.79 | 36.50 |
Spiking noise removal platform based on CORDIC_HR model | 20.3 | 34.1 | 9.2 | 36.8 |
