Denoising Malaria Color Images using Non-Local Means (NLM) Filter Based on Region of Interest
Wikan Tyassari, Siti Nurul Aqmariah Mohd Kanafiah, Yessi Jusman, Zeehaida Mohamed · 2024
Malaria is a life-threatening infectious disease spreading rapidly in the human body. People infected with the malaria parasite should receive immediate handling within 24 hours. Artificial intelligence (AI) assists medical experts in diagnosing malaria effectively and quickly. However, reliable classification necessitates image processing. A non-local means (NLM) filter was developed to reduce the noise in the images based on the four levels of region of interest (ROI). At low-level noise, NLM filtering using ROI outperformed NLM filtering using the average. NLM reduced noise based on the pixel of ROI. The highest performance occurred at an ROI of 75x75 pixels, with 38.40 ± 1.80 dB for PSNR, 36.27 ± 2.61 dB for SNR, 10.64 ±7.72 for MSE, and 0.94 ± 0.04 dB for similarity index. The filter depicted the greatest result in reducing Gaussian noise.