Medical Image De-Noising Using Discrete Wavelet Transform and Threshold Filter

Abstract

Medical imaging plays a critical role in diagnosis and treatment planning. However, during acquisition and transmission, medical images such as MRI, CT, or X-rays are often corrupted by noise, which may degrade the image quality and affect diagnostic accuracy. Effective de-noising is therefore essential to preserve important structural details while removing noise artifacts. This research presents an efficient de-noising method for medical images using Discrete Wavelet Transform (DWT) combined with threshold-based filtering. The DWT is employed to decompose the image into various subbands representing different frequency components. Since noise primarily affects the high-frequency subbands, appropriate thresholding techniques are applied to suppress these components while retaining the significant low-frequency details. The inverse DWT is then used to reconstruct the noise-reduced image. Performance metrics like Peak Signal-to-Noise Ratio (PSNR) and mean square error (MSE) are used to evaluate the quality of de-noised images.

Experimental results on standard medical image datasets demonstrate that the proposed DWT and threshold-based method effectively removes noise while preserving anatomical features. This method can be highly beneficial for pre-processing in automated diagnostic systems, improving clinical accuracy and reducing false diagnoses.