| 杨碧凝,刘宇翔,陈辛元,等.利用噪声等价图像和深度学习方法对低剂量CT降噪[J].中华放射医学与防护杂志,2022,42(5):355-360.Yang Bining,Liu Yuxiang,Chen Xinyuan,et al.Noise reduction in low-dose computed tomography with noise equivalent image and deep learning[J].Chin J Radiol Med Prot,2022,42(5):355-360 |
| 利用噪声等价图像和深度学习方法对低剂量CT降噪 |
| Noise reduction in low-dose computed tomography with noise equivalent image and deep learning |
| 投稿时间：2021-12-28 |
| DOI：10.3760/cma.j.cn112271-20211228-00499 |
| 中文关键词: 低剂量CT 噪声等价图像 深度学习 降噪 放疗模拟定位 |
| 英文关键词:Low-dose CT Noise equivalent image Deep learning Noise reduction Radiotherapy simulation |
| 基金项目:国家自然科学基金(11975313,12175312);北京科技新星计划(Z201100006820058) |
|
| 摘要点击次数: 3453 |
| 全文下载次数: 1276 |
| 中文摘要: |
| 目的 研究在常规剂量扫描情况下模拟低剂量CT图像的方法,以此生成训练数据集中与常规剂量CT具有对应关系的低剂量CT图像,并建立深度学习模型,用于低剂量CT图像的降噪。方法 使用Philip Brilliance CT Big Bore模拟定位机,其不同算法重建的CT图像具有不同的噪声水平,其中iDose4算法噪声较大,而全模型迭代重建技术(iterative model reconstruction,IMR)噪声较小。提出一种以等价噪声水平重建图像替代低剂量CT图像的方法。常规剂量和低剂量CT的曝光量分别采用250和35 mAs。分别扫描CTP712均匀模块,用IMR算法重建低剂量CT图像,用不同降噪水平的iDose4算法重建常规剂量CT图像,并根据噪声分布从中找出低剂量CT的噪声等价图像。随后,用常规剂量和噪声等价CT图像配对训练循环一致性生成对抗网络(cycle-consistent adversarial networks,CycleGAN),使用模体测试该方法对真实低剂量CT噪声的改善程度。结果 用iDose4 level 1重建的常规剂量CT图像可替代IMR重建的低剂量CT图像。低剂量扫描可降低86%的辐射剂量。使用CycleGAN模型降噪后,对于均匀模块,降噪幅度为45%;对于CIRS-SBRT 038模体的脑、脊髓和骨等处,噪声值分别降低了50%,13%和7%。结论 等价噪声水平重建图像可用于替代低剂量CT图像训练深度学习网络,在避免受检者受照剂量增加的同时,减少图像噪声,提高图像质量。 |
| 英文摘要: |
| Objective To investigate the method of simulating low-dose CT (LDCT) images using routine dose level scanning mode to generate LDCT images with correspondence to the routine dose CT (RDCT) images in the training sets for deep learning model, which would be used for LDCT noise reduction.Methods The CT images reconstructed by different algorithms in Philips CT Big Core had different noise levels, where the noise was larger with iDose4 algorithm and lower with IMR(knowledge-based iterative model reconstruction)algorithm. A new method of replacing LDCT image with noise equivalent reconstructed image was proposed. The uniform module of CTP712 was scanned with the exposure of 250mAs for RDCT, 35mAs for LDCT. The images were reconstructed using IMR algorithm for LDCT images and iDose4 algorithm at multiple noise reduction levels for RDCT images, respectively. The noise distribution of each image set was analyzed to find the noise equivalent images of LDCT. Then, RDCT images, those selected images were used for training cycle-consistent adversarial networks (CycleGAN)model, and the noise reduction ability of the proposed method on real LDCT images of phantom was tested.Results The RDCT images generated with iDose4 level 1 could substitute the LDCT images reconstructed with IMR algorithm. The radiation dose was reduced by 86% in low dose scanning. Using CycleGAN model, the noise reduction degree was 45% for uniform module, and 50%, 13%, 7% for CIRS-SBRT 038 phantom in the specific regions of brain, spinal cord, bone, respectively.Conclusions Equivalent noise level reconstructed images could potentially serve as the alternative of LDCT images for deep learning network training to avoid additional radiation dose. The generated CT images had substantially reduced noise relative to that of LDCT. |
| HTML 查看全文 查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|