Irregular feature enhancer for low-dose CT denoising

• Published in: Multimedia systems Vol. 30; no. 6
• Main Authors: Deng, Jiehang, Hu, Zihang, He, Jinwen, Liu, Jiaxin, Qiao, Guoqing, Gu, Guosheng, Weng, Shaowei
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.12.2024; Springer Nature B.V
• Subjects: Ablation; Aspect ratio; Asymmetry; Computed tomography; Computer Communication Networks; Computer Graphics; Computer Science; Convolution; Cryptology; Data Storage Representation; Deep learning; Feature extraction; Human tissues; Medical imaging; Modules; Multimedia Information Systems; Noise reduction; Operating Systems; Regular Paper; Low-dose; Hybrid loss; Denoising; SASCM
• ISSN: 0942-4962; 1432-1882
• DOI: 10.1007/s00530-024-01575-7

So far, deep learning-based networks have been widely applied in Low-Dose Computed Tomography (LDCT) image denoising. However, they usually adopt symmetric convolution to achieve regular feature extraction, but cannot effectively extract irregular features. Therefore, in this paper, an Irregular Feature Enhancer (IFE) focusing on effectively extracting irregular features is proposed by combining Symmetric-Asymmetric-Synergy Convolution Module (SASCM) with a hybrid loss module. The shape, size and aspect ratio of human tissues and lesions are irregular, whose features are difficult for symmetric square convolution to extract. Rather than simply stacking symmetric convolution layers used in traditional deep learning-based networks, the SASCM with certain combination order of symmetric and asymmetric convolutional layers is devised to extract the irregular features. To the best of our knowledge, the IFE is the first work to propose the hybrid loss combining MSE, multi-scale perception loss and gradient loss, and apply asymmetric convolution in the field of LDCT denoising. The ablation experiments demonstrate the effectiveness and feasibility of SASCM and the hybrid loss. The quantitative experimental results also show that in comparison with several related LDCT denoising methods, the proposed IFE performs the best in terms of PSNR and SSIM. Furthermore, it can be observed from the qualitative visualization that the proposed IFE can recover the best image detail structure information among the compared methods.