深度监督的基于变压器的噪声感知网络用于低剂量的PET,在不同的计数级别中消除噪声
Mohammad Saber Azimi1, Vahid Felfelian1, Navid Zeraatkar2
1Department of Medical Radiation Engineering, Shahid Beheshti University, Tehran, Iran.
Computers in biology and medicine
|July 9, 2025
概括
一个新的基于Swin变压器的统一噪声感知网络 (ST-UNN) 有效地消除了不同噪声级别的低剂量PET图像. 这种深度学习方法可以提高图像质量和诊断可靠性,而不需要多个模型.
科学领域:
- 医疗成像医学成像
- 深度学习 (Deep Learning) 是一种深度学习.
- 放射学 放射学是一门学科.
背景情况:
- 降低辐射剂量在正子发射断层扫描 (PET) 成像中对于最大限度地降低癌症风险至关重要.
- 低剂量PET成像通常会导致噪音增加和图像质量降低,影响诊断准确度.
- 现有的深度学习无声化方法很难在PET扫描中常见的各种噪声水平上进行概括.
研究的目的:
- 开发一个统一的深度学习网络,能够处理低剂量PET成像中的不同噪音水平.
- 从低剂量采集中重建高质量的PET图像,使用基于Swin变压器的方法.
- 克服现有模型的局限性,这些模型需要针对特定噪声水平的培训.
主要方法:
- 一个基于Swin变压器的噪声感知网络 (ST-UNN) 被开发出来,集成了多个子网络,用于1%至10%的噪声水平.
- 采用了自适应权重机制,以动态组合子网络的输出,以便有效地消除噪音.
- 该模型在头部和部区域的PET/CT数据集上进行训练和验证,评估SSIM,PSNR,SUV偏差和RMSE的性能.
主要成果:
- 与传统网络相比,ST-UNN的性能优越,特别是在超低剂量场景 (1%计数水平) 中.
- 在1%的计数水平上,ST-UNN的PSNR达到34.77,RMSE为0.05,SSIM为0.97,SUV偏差最小.
- 网络在所有测试的噪声级别中保持了高性能和低误差,表明强大的概括性和诊断完整性.
结论:
- ST-UNN提供了一个可扩展的,基于变压器的解决方案,用于消除低剂量的PET图像.
- 副网络的动态集成有效地解决了噪声变化,显著提高了图像质量.
- 这种方法提升了低剂量和动态PET成像的能力,提高了诊断可靠性.
相关概念视频
01:20Downsampling
260
When considering a sampled sequence with zero values between sampling instants, one can replace it by taking every N-th value of the sequence. At these integer multiples of N, the original and sampled sequences coincide. This process, known as decimation, involves extracting every N-th sample from a sequence, thereby creating a more efficient sequence.
The Fourier transform of the decimated sequence reveals a combination of scaled and shifted versions of the original spectrum. This...
The Fourier transform of the decimated sequence reveals a combination of scaled and shifted versions of the original spectrum. This...
260
01:20Deconvolution
262
Deconvolution, also known as inverse filtering, is the process of extracting the impulse response from known input and output signals. This technique is vital in scenarios where the system's characteristics are unknown, and they must be inferred from the observable signals.
Deconvolution involves several mathematical techniques to derive the impulse response. One common approach is polynomial division. In this method, the input and output sequences are treated as coefficients of...
Deconvolution involves several mathematical techniques to derive the impulse response. One common approach is polynomial division. In this method, the input and output sequences are treated as coefficients of...
262