Related Experiment Video
Updated: Nov 29, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
3.2K
Denoising Medical Images Using Machine Learning, Deep Learning Approaches: A Survey.
Ali Arshaghi1, Mohsen Ashourian2, Leila Ghabeli1
1Department of Electrical Engineering, Central Tehran Branch, Islamic Azad University, Tehran, Iran.
Current Medical Imaging
|November 20, 2020
Summary
This study compares medical image denoising techniques, finding that deep convolutional neural networks (CNNs) outperform traditional filters in reducing noise like Gaussian and salt-and-pepper. The CNN method achieved superior performance metrics.
Area of Science:
- Medical Imaging
- Image Processing
- Computational Science
Background:
- Medical image quality is crucial for accurate diagnosis.
- Various denoising methods exist, including Wavelet Transform, CNNs, and linear/non-linear filters.
- Noise degrades image quality, necessitating effective denoising strategies.
Purpose of the Study:
- To evaluate and compare the performance of different medical image denoising techniques.
- To investigate the efficacy of a modified median filter algorithm.
- To analyze the application of Wavelet Transform, Non-local means (NLM), and deep convolutional neural network (Dn-CNN) for noise reduction.
Main Methods:
- Modified median filter algorithm.
- Application of Wavelet Transform, Non-local means (NLM), and deep convolutional neural network (Dn-CNN).
- Testing denoising performance using Signal-to-Noise Ratio (SNR), Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM), and Mean Squared Error (MSE).
Main Results:
- Deep convolutional neural network (CNN) methods demonstrated higher PSNR values compared to traditional filters.
- CNNs outperformed Adaptive Wiener filter, Median filter, Adaptive Median filter, and Wiener filter.
- Simulation results for image denoising were presented, highlighting performance differences.
Conclusions:
- Deep convolutional neural networks (CNNs) offer superior performance for medical image denoising.
- The study provides a comparative analysis of various denoising techniques.
- Quantitative metrics (SSIM, PSNR, MSE) confirm the effectiveness of advanced methods like CNNs.
