Histogram-Based Hybrid Wavelet-Curvelet Denoising for Medical Images

Abstract

Medical images often contain noise that degrades image quality and obscures important diagnostic details. This paper proposes a histogram-based hybrid denoising method that integrates Wavelet Transform and Curvelet Transform to enhance medical image clarity. The wavelet transform is utilized for efficient noise suppression in high-frequency components, while the curvelet transform effectively preserves edge information and directional features. A histogrambased adaptive fusion strategy is employed to combine the outputs of both transforms by analyzing intensity distribution and local image characteristics. The proposed method is implemented using MATLAB and evaluated on various medical images. Performance is assessed using quantitative metrics such as Peak Signal-toNoise Ratio (PSNR) and Structural Similarity Index (SSIM). Experimental results demonstrate that the proposed hybrid approach outperforms conventional denoising techniques by achieving higher PSNR and SSIM values, along with improved visual quality. The enhanced images retain critical structural details, making the method suitable for reliable pre-diagnostic analysis and medical image interpretation.