A new hybrid image denoising algorithm using adaptive and modified decision-based filters for enhanced image quality.

Area of Science:

• Digital Image Processing
• Computer Vision
• Signal Processing

Background:

• Traditional image denoising methods struggle with computational complexity, over-smoothing, and preserving critical details like edges.
• Effective noise reduction is crucial for recovering visual quality and structural integrity in digital images.

Purpose of the Study:

• To introduce a hybrid denoising algorithm combining Adaptive Median Filter (AMF) and Modified Decision-Based Median Filter (MDBMF).
• To address limitations of traditional methods by preserving edges and reducing noise effectively.

Main Methods:

• A hybrid denoising algorithm integrating AMF for dynamic window adjustment and MDBMF for selective pixel recovery.
• Testing on nine benchmark images, including standard and medical datasets (Chest, Liver) with varying noise densities (10-90%).

Main Results:

• The hybrid approach significantly outperforms state-of-the-art methods in subjective and objective analyses.
• Quantitative metrics show improvements: up to 2.34 dB in PSNR, over 20% in IEF, up to 15% in MSE, and 0.07 in SSIM.
• FOM and VIF metrics reached 0.68 and 0.61, respectively, demonstrating superior performance.

Conclusions:

• The proposed hybrid denoising algorithm offers superior performance in noise reduction while preserving image details, especially edges.
• This method provides a significant advancement over existing denoising techniques for both standard and medical imaging applications.