Scholarlink

<p>Medical imaging plays a key role in diagnosing and treating diseases. However, images often contain noise, which can make it hard for doctors and computers to see important details. This noise can come from the imaging device, the environment, or errors during data transmission. Removing noise from medical images is essential to improve their quality and help in accurate diagnosis. This paper reviews different methods used for image denoising, from traditional techniques like filtering and wavelet transforms to modern deep learning approaches. While classical methods such as anisotropic diffusion and non-local means filtering work well, they sometimes struggle to preserve fine details in images. On the other hand, deep learning models, including convolutional neural networks (CNNs) and generative adversarial networks (GANs), have shown great success in learning noise patterns and improving image clarity. Hybrid approaches, which combine deep learning with traditional filtering techniques, are also gaining attention for their effectiveness across different types of medical images, such as MRI, CT, ultrasound, and PET scans. Despite these advancements, challenges remain, including the need for large amounts of high-quality data, the difficulty of explaining how deep learning models work, and the demand for faster processing. This paper highlights these issues and discusses possible future directions in medical image denoising. By improving image quality, these techniques can enhance medical diagnosis, leading to better patient care. The objective of this study is to provide a comprehensive examination of several denoising methodologies employed in medical imaging modalities, Magnetic Resonance (MR), Computed Tomography (CT), Positron Emission Tomography (PET) images and other modalities.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Advances in Deep Learning and Filtering Models for Medical Image Denoising: A Review of Current and Future Trends

  • 期刊论文

  • 发表日期：

    2026年2月4日

查看全文
  • Kirti Pathak,
  • Ashish Kumar Bhandari

摘要

Medical imaging plays a key role in diagnosing and treating diseases. However, images often contain noise, which can make it hard for doctors and computers to see important details. This noise can come from the imaging device, the environment, or errors during data transmission. Removing noise from medical images is essential to improve their quality and help in accurate diagnosis. This paper reviews different methods used for image denoising, from traditional techniques like filtering and wavelet transforms to modern deep learning approaches. While classical methods such as anisotropic diffusion and non-local means filtering work well, they sometimes struggle to preserve fine details in images. On the other hand, deep learning models, including convolutional neural networks (CNNs) and generative adversarial networks (GANs), have shown great success in learning noise patterns and improving image clarity. Hybrid approaches, which combine deep learning with traditional filtering techniques, are also gaining attention for their effectiveness across different types of medical images, such as MRI, CT, ultrasound, and PET scans. Despite these advancements, challenges remain, including the need for large amounts of high-quality data, the difficulty of explaining how deep learning models work, and the demand for faster processing. This paper highlights these issues and discusses possible future directions in medical image denoising. By improving image quality, these techniques can enhance medical diagnosis, leading to better patient care. The objective of this study is to provide a comprehensive examination of several denoising methodologies employed in medical imaging modalities, Magnetic Resonance (MR), Computed Tomography (CT), Positron Emission Tomography (PET) images and other modalities.