| 基于双域自适应网络的岩矿样工业CT图像金属伪影校正算法研究 |
| |
| 引用本文： | 左顺吉,冯鹏,黄盼,严笙豪,何鹏,魏彪.基于双域自适应网络的岩矿样工业CT图像金属伪影校正算法研究[J].CT理论与应用研究,2022,31(6):783-792. |
| |
| 作者姓名： | 左顺吉 冯鹏 黄盼 严笙豪 何鹏 魏彪 |
| |
| 作者单位： | 1.重庆大学光电技术及系统教育部重点实验室, 重庆400044 |
| |
| 基金项目： | 科技部国家重点研发计划（重点锂、铍成矿带成矿规律与预测评价研究与综合（2019YFC0605203））；重庆市科委基础研究与前沿探索专项（自然科学基金）（面向小样本的管激发X射线荧光CT成像关键技术研究（cstc2020jcyj-msxmX0553））；重庆市科委技术创新与应用发展专项（轨道交通智慧化车站研究及应用（cstc2021jscx-gksbX0056））。 |
| |
| 摘 要： | 岩矿样中包含有大量高密度金属物质，致使其在工业CT图像上产生了金属伪影，严重影响岩矿样参数分析的准确性。为抑制岩矿样CT图像的金属伪影，本文提出一种基于双域自适应网络的岩矿样CT图像金属伪影校正算法（DDA-CNN-MAR），将含有金属伪影的CT图像分别通过投影域网络和图像域网络进行金属伪影的抑制，自适应融合双域处理结果，实现由含伪影图像到无伪影图像的端到端映射。该算法以残差编解码网络模型（RED-CNN）为基础，易于提取特征并恢复图像细节；双域结构可自适应调整投影域（伪影抑制）和图像域（细节修复）的权重，借以获得最优的校正结果。研究结果表明，较之于RED-CNN-MAR，经过DDA-CNN-MAR方法校正的图像，MSE减小2.570，而PSNR和SSIM则分别提高1.218 dB和0.018，有效提升岩矿样CT成像的图像质量。
|
| 关 键 词： | 工业CT 深度学习 校正算法 金属岩矿样 金属伪影 |
| 收稿时间： | 2022-03-11 |
Metal Artifact Reduction Algorithm for CT Images of Rock and Mineral Samples Based on Dual-domain Adaptive Network |
| |
| Institution: | 1.Key Laboratory of Optoelectronic Technology and System, Ministry of Education, Chongqing University, Chongqing 400044, China2.Engineering Research Center of Industrial CT Nondestructive Testing, Ministry of Education, Chongqing University, Chongqing 400044, China |
| |
| Abstract: | Rock and mineral samples contain a large amount of high-density metal substances, which often lead to metal artifacts in CT images and seriously affect the parameters analysis accuracy of rock and mineral samples. In order to improve the quality of CT images and suppress metal artifacts, in this paper we propose a dual-domain adaptive network-based metal artifact reduction algorithm for CT images of rock samples (DDA-CNN-MAR). The metal artifacts are suppressed by the projection domain network and the image domain network successively, and the dual domain processing results are adaptively fused to realize the end-to-end mapping from images with artifacts to artifact-free images. The algorithm is based on the residual encoder-decoder network model (RED-CNN), which is easy to extract features and restore image details. The dual-domain structure can adaptively adjust the weights of the projection domain (artifact suppression) and image domain (detail inpainting) to obtain optimal reduction results. The experimental results show that, compared with the RED-CNN metal artifact denoising method in the image domain, the MSE of the image corrected by the DDA-CNN-MAR method is reduced by 2.570, while the PSNR and SSIM are increased by 1.218 dB and 0.018 respectively, which effectively improves the CT image quality. |
| |
| Keywords: | |
|
| 点击此处可从《CT理论与应用研究》浏览原始摘要信息 |
| 点击此处可从《CT理论与应用研究》下载免费的PDF全文 |
