Published July 4, 2025 | Version v1
Figure Open

CNLNet Enhancing MRI Brain Image Denoising Using a Convolutional Neural Network with Integrated Non-Local Means Layer - Figure 3

  • 1. Karunya Institute of Technology and Sciences
  • 2. Samara State Medical University

Description

The  dataset  consists  of  3,072  MRI  brain  images  sourced  from a  publicly  accessible  Kaggle 
dataset  at  https://www.kaggle.com/datasets/abhranta/brain-tumor-detection-mri.  Images  were 
resized  to  a  uniform  resolution  of  128×128×3  and  normalised  to  the  range  [0,  1]  for  consistent 
scaling  and  improved  convergence  during  training.  To  construct  noisy-clean  image  pairs,  3  SD  of 
Gaussian  noise  (0.1,  0.3,  and  0.5)  were  applied.  Figure  3  displays  sample  images  of  clean  and  noisy 
images with SD of 0.1, 0.3, and 0.5, respectively.

Notes

Files

CNLNet Enhancing MRI Brain Image Denoising Using a Convolutional Neural Network with Integrated Non-Local Means Layer - Figure 3.png