Scholarlink

Background <p>Deep learning (DL) is becoming increasingly popular for analyzing magnetic resonance imaging (MRI), particularly in neuroimaging. Image denoising is a crucial preprocessing step in medical image analysis. This prospective study was conducted in 28 patients who underwent brain MRI. We aimed to investigate how DL improves the signal-to-noise ratio (SNR) of the T2-weighted image in&#xa0;brain imaging.</p> Results <p>SNR was statistically significantly higher in DL image analysis compared to original image analysis. This improvement in the SNR ratio in DL images is explained by the significant decrease in noise (N) without a significant change in image signal (S). There was a statistically significant consistency and absolute agreement between the original and DL image analysis on assessing SNR.</p> Conclusions <p>Deep learning-based denoising techniques can significantly improve the quality of T2-weighted MRI images by systematic shift in SNR values.</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

Deep learning-based image denoising can improve the signal-to-noise ratio of the T2-weighted image in brain imaging

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月17日

查看全文
  • Somyya M. Ghanim,
  • Sabry Elmogy,
  • Gehad A. Saleh

摘要

Background

Deep learning (DL) is becoming increasingly popular for analyzing magnetic resonance imaging (MRI), particularly in neuroimaging. Image denoising is a crucial preprocessing step in medical image analysis. This prospective study was conducted in 28 patients who underwent brain MRI. We aimed to investigate how DL improves the signal-to-noise ratio (SNR) of the T2-weighted image in brain imaging.

Results

SNR was statistically significantly higher in DL image analysis compared to original image analysis. This improvement in the SNR ratio in DL images is explained by the significant decrease in noise (N) without a significant change in image signal (S). There was a statistically significant consistency and absolute agreement between the original and DL image analysis on assessing SNR.

Conclusions

Deep learning-based denoising techniques can significantly improve the quality of T2-weighted MRI images by systematic shift in SNR values.