TY  - JOUR
T1  - Development and Verification of Novel Black and White Area Preserving
Salt and Pepper Noise Removal Image Processing Design
AU - Jain, Abhishek AU - Gupta, Richa 
JO  - Journal of Engineering and Applied Sciences
VL  - 14
IS  - 6
SP  - 1828
EP  - 1839
PY  - 2019
DA  - 2001/08/19
SN  - 1816-949x
DO  - jeasci.2019.1828.1839
UR  - https://makhillpublications.co/view-article.php?doi=jeasci.2019.1828.1839
KW  - Salt and pepper noise
KW  -grayscale images
KW  -functional verification
KW  -SystemVerilog
KW  -Universal
Verification Methodology (UVM)
KW  -Mean Square Error (MSE)
KW  -Peak Signal to Noise Ratio (PSNR)
KW  -Structural Similarity Index (SSIM)
AB  - In semiconductor industry, image processing algorithms are developed and evaluated using software
models (‘C’/MATLAB) before actual implementation of RTL. After the evaluation of the algorithm, software
models are used as a golden reference model for the RTL development of the image processing algorithms. In
this study, we are describing the novel black and white area preserving salt and pepper noise removal algorithm
and its RTL implementation using Verilog language. SystemVerilog UVM based verification environment of salt
and pepper noise removal RTL design is developed. Quality of different image denoising algorithms is
quantitatively measured by different parameters, namely Mean Square Error (MSE), Peak Signal to Noise Ratio
(PSNR), Structural Similarity Index (SSIM) and Multi Scale Structural Similarity Index (MS-SSIM). Time
complexity is measured by Big O notation and stopwatch timer. The main motivation behind this work is to
propose best efficient black and white area preserving salt and pepper noise removal algorithm, its RTL
implementation and development of efficient functional verification framework of salt and pepper noise removal
RTL design. The proposed algorithm gives better MSE, PSNR, SSIM and MS-SSIM results. Time taken by
proposed algorithm is not less than other existing algorithms but it is acceptable as Big O notation is same for
other algorithms.
ER  -