FSformer: A combined frequency separation network and transformer for LDCT denoising

被引：4
作者：
Kang J. [1 ,2 ]
Liu Y. [1 ,2 ]
Zhang P. [1 ,2 ]
Guo N. [1 ,2 ]
Wang L. [1 ,2 ]
Du Y. [1 ,2 ]
Gui Z. [1 ,2 ]
机构：
[1] State Key Laboratory of Dynamic Testing Technology, North University of China, Taiyuan
[2] School of Information and Communication Engineering, North University of China, Taiyuan
来源：
Computers in Biology and Medicine | 2024年 / 173卷
关键词：
Frequency separation network; Low-dose computed tomography (LDCT); Noise/artifact removing; Transformer;
D O I：
10.1016/j.compbiomed.2024.108378
中图分类号：
学科分类号：
摘要：
Low-dose computed tomography (LDCT) has been widely concerned in the field of medical imaging because of its low radiation hazard to humans. However, under low-dose radiation scenarios, a large amount of noise/artifacts are present in the reconstructed image, which reduces the clarity of the image and is not conducive to diagnosis. To improve the LDCT image quality, we proposed a combined frequency separation network and Transformer (FSformer) for LDCT denoising. Firstly, FSformer decomposes the LDCT images into low-frequency images and multi-layer high-frequency images by frequency separation blocks. Then, the low-frequency components are fused with the high-frequency components of different layers to remove the noise in the high-frequency components with the help of the potential texture of low-frequency parts. Next, the estimated noise images can be obtained by using Transformer stage in the frequency aggregation denoising block. Finally, they are fed into the reconstruction prediction block to obtain improved quality images. In addition, a compound loss function with frequency loss and Charbonnier loss is used to guide the training of the network. The performance of FSformer has been validated and evaluated on AAPM Mayo dataset, real Piglet dataset and clinical dataset. Compared with previous representative models in different architectures, FSformer achieves the optimal metrics with PSNR of 33.7714 dB and SSIM of 0.9254 on Mayo dataset, the testing time is 1.825 s. The experimental results show that FSformer is a state-of-the-art (SOTA) model with noise/artifact suppression and texture/organization preservation. Moreover, the model has certain robustness and can effectively improve LDCT image quality. © 2024 Elsevier Ltd
引用
收藏
相关论文
共 46 条
[1]  
Brenner D.J., Hall E.J., Computed tomography—an increasing source of radiation exposure, N. Engl. J. Med., 357, pp. 2277-2284, (2007)
[2]  
Lu Z., Xia W., Huang Y., Et al., M3NAS: multi-scale and multi-level memory-efficient neural architecture search for low-dose CT denoising, IEEE Trans. Med. Imag., 42, pp. 850-863, (2023)
[3]  
Balda M., Hornegger J., Heismann B., Et al., Ray contribution masks for structure adaptive sinogram filtering, IEEE Trans. Med. Imag., 31, pp. 1228-1239, (2012)
[4]  
Manduca A., Yu L., Trzasko J.D., Et al., Projection space denoising with bilateral filtering and CT noise modeling for dose reduction in CT, Med. Phys., 36, pp. 4911-4919, (2009)
[5]  
Wang J., Li T., Lu H., Et al., Penalized weighted least-squares approach to sinogram noise reduction and image reconstruction for low-dose X-ray computed tomography, IEEE Trans. Med. Imag., 25, pp. 1272-1283, (2006)
[6]  
Chen G., Et al., AirNet: fused analytical and iterative reconstruction with deep neural network regularization for sparse-data CT, Med. Phys., 47, pp. 2916-2930, (2020)
[7]  
Niu S., Gao Y., Bian Z., Et al., Sparse-view x-ray CT reconstruction via total generalized variation regularization, Phys. Med. Biol., 59, pp. 2997-3017, (2014)
[8]  
Bai T., Mou X., Xu Q., Et al., Low-dose CT reconstruction based on multiscale dictionary, SPIE Medical Imaging, (2013)
[9]  
Kang J., Liu Y., Shu H., Et al., Edge protection and global attention mechanism densely connected convolutional network for LDCT denoising, circuits, System Signal Process, (2023)
[10]  
Yan R., Liu Y., Liu Y., Wang L., Zhao R., Bai Y., Gui Z., Image denoising for low-dose CT via convolutional dictionary learning and neural network, IEEE Transactions on Computational Imaging, 9, pp. 83-93, (2023)
12345