TY  - JOUR
AU  - Katta, Swapna
AU  - Garg, Deepak
PY  - 2025
DA  - 2025/09/22
TI  - CT Image Denoising using Discrete Wavelet Transform
JO  - Biomedical Informatics and Smart Healthcare
T2  - Biomedical Informatics and Smart Healthcare
JF  - Biomedical Informatics and Smart Healthcare
VL  - 1
IS  - 2
SP  - 44
EP  - 51
DO  - 10.62762/BISH.2025.874472
UR  - https://www.icck.org/article/abs/BISH.2025.874472
KW  - CT image
KW  - denoising Gaussian noise
KW  - DWT
KW  - Transform domain
AB  - Low Dose Computed Tomography (LDCT) scan is modern medical imaging diagnostic technique that provides a detailed projection of internal human body tissue level structures. Even though the LDCT image quality is compromised by Gaussian-noise, which can be generated during image acquisition, this compromises the accurate diagnostic precision. The effective denoising is required to improve image quality in LDCT images. This study demonstrates that the Discrete Wavelet Transform(DWT) method shows better results, both quantitatively and visually, under varying noise intensities ($\sigma=10,20,30,$ and $40$). The DWT method decomposes the image to multiresolution subbands (approximation, and detail) to provide localized analysis of structural patterns. The thresholding method is applied to the detail (noisy) coefficients and then reconstructs the refined image from these denoised coefficients. The DWT method achieved superior noise suppression while preserving edge information. The quantitative analysis among various methods, including PCA, MSVD, DCT, and DWT, consistently shows superior results, achieving a higher PSNR of $33.85$ dB, SNR of $28.50$ dB, and SSIM of $0.7194$ at a noise level $\sigma =10$. Among all denoising methods, the DWT is a powerful and consistent method in image processing to enhance image quality in LDCT images.
SN  - 3068-5524
PB  - Institute of Central Computation and Knowledge
LA  - English
ER  - 

@article{Katta2025CT,
  author = {Swapna Katta and Deepak Garg},
  title = {CT Image Denoising using Discrete Wavelet Transform},
  journal = {Biomedical Informatics and Smart Healthcare},
  year = {2025},
  volume = {1},
  number = {2},
  pages = {44-51},
  doi = {10.62762/BISH.2025.874472},
  url = {https://www.icck.org/article/abs/BISH.2025.874472},
  abstract = {Low Dose Computed Tomography (LDCT) scan is modern medical imaging diagnostic technique that provides a detailed projection of internal human body tissue level structures. Even though the LDCT image quality is compromised by Gaussian-noise, which can be generated during image acquisition, this compromises the accurate diagnostic precision. The effective denoising is required to improve image quality in LDCT images. This study demonstrates that the Discrete Wavelet Transform(DWT) method shows better results, both quantitatively and visually, under varying noise intensities (\$\sigma=10,20,30,\$ and \$40\$). The DWT method decomposes the image to multiresolution subbands (approximation, and detail) to provide localized analysis of structural patterns. The thresholding method is applied to the detail (noisy) coefficients and then reconstructs the refined image from these denoised coefficients. The DWT method achieved superior noise suppression while preserving edge information. The quantitative analysis among various methods, including PCA, MSVD, DCT, and DWT, consistently shows superior results, achieving a higher PSNR of \$33.85\$ dB, SNR of \$28.50\$ dB, and SSIM of \$0.7194\$ at a noise level \$\sigma =10\$. Among all denoising methods, the DWT is a powerful and consistent method in image processing to enhance image quality in LDCT images.},
  keywords = {CT image, denoising Gaussian noise, DWT, Transform domain},
  issn = {3068-5524},
  publisher = {Institute of Central Computation and Knowledge}
}