Scholarlink

Background <p>We used the deep learning-based reconstruction algorithm to reduce the scan time for brain T2-weighted images (T2WI) with reduction of image noise and preservation of image quality. The current prospective study included 22 cases; they were subjected to brain magnetic resonance. The aim was to investigate how much could deep learning reduce acquisition time of T2WI in brain imaging.</p> Results <p>There was a statistically significantly higher air noise (N) in conventional image analysis versus DL image analysis, but no statistically significant difference in image signal (S). Signal-to-noise ratio (SNR) was statistically significantly higher in DL image analysis versus conventional image analysis. This showed that the improvement in SNR was explained by significant decrease in N without significant change in S.</p> Conclusions <p>We concluded that deep learning-based reconstruction provides advantages of maintaining optimum image quality, reduction of image noise and obtaining the images in short time while applied during acquisition of brain magnetic resonance imaging (MRI) by T2WI. Yet, the diagnostic value of this technique to be verified, large-scale clinical studies should be conducted on this issue.</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 can reduce acquisition time of T2-weighted image in brain imaging

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年2月19日

查看全文
  • Eman Hassan El-Saeed Abou-ELMagd,
  • Sabry Alameldin Elmogy,
  • Dina Gamal Abdelzaher

摘要

Background

We used the deep learning-based reconstruction algorithm to reduce the scan time for brain T2-weighted images (T2WI) with reduction of image noise and preservation of image quality. The current prospective study included 22 cases; they were subjected to brain magnetic resonance. The aim was to investigate how much could deep learning reduce acquisition time of T2WI in brain imaging.

Results

There was a statistically significantly higher air noise (N) in conventional image analysis versus DL image analysis, but no statistically significant difference in image signal (S). Signal-to-noise ratio (SNR) was statistically significantly higher in DL image analysis versus conventional image analysis. This showed that the improvement in SNR was explained by significant decrease in N without significant change in S.

Conclusions

We concluded that deep learning-based reconstruction provides advantages of maintaining optimum image quality, reduction of image noise and obtaining the images in short time while applied during acquisition of brain magnetic resonance imaging (MRI) by T2WI. Yet, the diagnostic value of this technique to be verified, large-scale clinical studies should be conducted on this issue.