New Real-Time High-Density Impulsive Noise Removal Method Applied to Medical Images.

Area of Science:

• Medical Imaging
• Image Processing
• Computer Vision

Background:

• High-density impulsive noise severely degrades medical image quality, complicating diagnosis.
• Existing noise elimination methods struggle with color information loss around corrupted pixels, leading to suboptimal restoration.

Purpose of the Study:

• To introduce a new real-time method for impulsive noise elimination in medical images.
• To address the challenge of restoring images with significant noise density and color information loss.

Main Methods:

• A novel Nested Filtering followed by a Morphological Operation (NFMO) technique is proposed for pixel replacement.
• Noise detection is performed using the Modified Laplacian Vector Median Filter (MLVMF).
• An optimized Field-Programmable Gate Array (FPGA) design is suggested for real-time implementation.

Main Results:

• NFMO achieved excellent restoration quality on images with up to 90% impulsive noise, demonstrated on the Lena image (PSNR 29.99 dB).
• For medical images under identical noise conditions, NFMO achieved an average PSNR of 31.62 dB and an average NCD of 0.10.
• Real-time processing of medical images was achieved in an average time of 23 milliseconds.

Conclusions:

• The proposed NFMO method offers superior performance for high-density impulsive noise elimination in medical images.
• NFMO effectively restores corrupted pixels and preserves local data, outperforming traditional replacement techniques.
• The method is suitable for real-time applications due to its efficient FPGA implementation.