Year: 2024
Type: article
Abstract: This paper introduces a novel approach for enhancing MRI images corrupted by impulse noise. Impulse noise often degrades image quality, especially in medical imaging applications where accurate representation is critical. The proposed method employs a custom filter to remove impulse noise while maintaining essential image features. The custom filter is designed by altering the average value of its surrounding pixels depending on whether the pixel is entirely black or completely white. The filter’s design is determined by the specifications of MRI images and noise patterns typically encountered in such data. The proposed custom filter exhibits the effective operation of the denoising technique and fully utilizes the benefits of the improved median filter in eliminating impulse noise. The results demonstrate an improvement in image quality based on various quality metrics. There is a 6% improvement in Peak Signal Noise Ratio, 0.5% in Mean Square Error, 1% in Structural Similarity Index Measure, 0.7% in Image Quality and 0.3% in Average Gradient compared to a conventional median filter. Experimental results demonstrate significant improvements in image quality and preservation of diagnostic information compared to existing methods. The proposed approach offers a promising solution for enhancing MRI images affected by impulse noise, thereby aiding in more accurate diagnosis and treatment planning in medical imaging.
Authors: Kalpana Patil, Varsha Bendre
Institution: Savitribai Phule Pune University
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4410617044