Enhanced Denoising of Cervical Pre-cancerous Cell Images through Advanced Color Filtering
Maryza Intan Rahmawati, Siti Marhainis Othman, Siti Nurul Aqmariah Mohd Kanafiah, Yessi Jusman, Anani Aila Mat Zin, Nur Syuhada Mohd Nafis · 2024
Some disturbances in images will significantly degrade their accuracy and quality. Various types of microscopes cause images to be contaminated by noise, uneven lighting, poor contrast, and blur. In pre-processing, noise filtering is performed to detect and eliminate noise in the images. Image data in this study contained several types of disturbing noise: Poisson noise, Speckle noise, and Gaussian noise. Noise filtering was conducted using Gaussian, Median, Wiener, Bilateral, and Local Laplacian filters with three window sizes. The values of Peak Signal-to-Noise Ratio (PSNR), Signal-toNoise Ratio (SNR), Structural Similarity Index Measure (SSIM), and Mean Squared Error (MSE) metrics were employed to assess the performance of several filters in every denoised image. The outcomes demonstrated that a Gaussian filter with a window size of 5 outperformed other filters in eliminating Gaussian noise and Speckle noise. Meanwhile, Poisson noise was efficiently removed using a Wiener filter with a window size of $7 \times 7$. Sometimes, a bigger or smaller window size could generate the greatest result.