|
|
自动检测MR图像中的伪影：基于改进的AMAResNet网络
|
Abstract:
本文提出了一种创新网络结构，基于改进ResNet网络的算法，针对传统方法对局部特征和多尺度信息捕捉不足的问题，引入分块机制、Squeeze-and-Excitation (SE)模块和金字塔模块，分块机制将输入图像划分为局部子块以增强细节感知；SE模块通过通道权重自适应优化关键特征表达；金字塔结构融合跨尺度特征以提升复杂伪影识别能力。为解决类别不平衡问题，提出交叉熵与Dice损失的加权联合损失函数(最优权重比α = 0.7/β = 0.3)，构建出一个全新的名为AMAResNet的伪影检测模型，从而实现了MRI伪影的自动检测。该方法是数据驱动的，且不需要特殊的硬件要求，使用空间域中MR图像的大小作为输入。本文在真实数据上评估了算法性能，结果显示，改进后的网络在不同类型数据上的准确性和泛化能力方面均有显著提升。所提出的方法有望在扫描完成后立即将MR图像标记为无伪影或有伪影状态，从而确保获取高质量的MR图像，这对于准确的医学诊断至关重要。实验结果表明，改进后的ResNet网络在自动检测MRI伪影方面表现出色，具备在临床实践中广泛应用的潜力，有助于提高MRI图像质量和诊断准确性。
This paper proposes an innovative network architecture based on an enhanced ResNet algorithm to address the limitations of traditional methods in capturing local features and multi-scale information. The framework introduces three key components: a partition mechanism that divides input images into local sub-regions to enhance detail perception, Squeeze-and-Excitation (SE) modules that adaptively optimize critical feature representations through channel-wise weighting, and a pyramid structure that integrates cross-scale features to improve complex artifact recognition. To resolve class imbalance issues, we design a weighted joint loss function combining cross-entropy and Dice Loss (optimal weight ratio α = 0.7/β = 0.3), ultimately constructing a novel artifact detection model named AMAResNet for automatic MRI artifact detection. This data-driven approach requires no specialized hardware and utilizes spatial domain MR image sizes as input. Evaluations of real-world datasets demonstrate that the enhanced network achieves significant improvements in both accuracy and generalization across different data types. The proposed method shows potential for immediate post-scan quality assessment by labeling MR images as artifact-free or artifact-affected, thereby ensuring high-quality medical imaging crucial for accurate diagnosis. Experimental results indicate that the improved ResNet architecture excels in automatic MRI artifact detection, exhibiting strong potential for clinical application to enhance both MRI image quality and diagnostic reliability.
| [1] | Zaitsev, M., Maclaren, J. and Herbst, M. (2015) Motion Artifacts in MRI: A Complex Problem with Many Partial Solutions. Journal of Magnetic Resonance Imaging, 42, 887-901. https://doi.org/10.1002/jmri.24850 |
| [2] | Maclaren, J., Herbst, M., Speck, O. and Zaitsev, M. (2012) Prospective Motion Correction in Brain Imaging: A Review. Magnetic Resonance in Medicine, 69, 621-636. https://doi.org/10.1002/mrm.24314 |
| [3] | Heiland, S. (2008) From a as in Aliasing to Z as in Zipper: Artifacts in MRI. Clinical Neuroradiology, 18, 25-36. https://doi.org/10.1007/s00062-008-8003-y |
| [4] | Wood, M.L. and Henkelman, R.M. (1985) MR Image Artifacts from Periodic Motion. Medical Physics, 12, 143-151. https://doi.org/10.1118/1.595782 |
| [5] | Meding, K., Loktyushin, A. and Hirsch, M. (2017) Automatic Detection of Motion Artifacts in MR Images Using CNNS. 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), New Orleans, 5-9 March 2017, 811-815. https://doi.org/10.1109/icassp.2017.7952268 |
| [6] | Küstner, T., Liebgott, A., Mauch, L., Martirosian, P., Bamberg, F., Nikolaou, K., et al. (2017) Automated Reference-Free Detection of Motion Artifacts in Magnetic Resonance Images. Magnetic Resonance Materials in Physics, Biology and Medicine, 31, 243-256. https://doi.org/10.1007/s10334-017-0650-z |
| [7] | Kustner, T., Jandt, M., Liebgott, A., Mauch, L., Martirosian, P., Bamberg, F., et al. (2018) Automatic Motion Artifact Detection for Whole-Body Magnetic Resonance Imaging. 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Calgary, 15-20 April 2018, 995-999. https://doi.org/10.1109/icassp.2018.8462414 |
| [8] | Oksuz, I., Clough, J., Ruijsink, B., Puyol-Antón, E., Bustin, A., Cruz, G., et al. (2019) Detection and Correction of Cardiac MRI Motion Artefacts during Reconstruction from K-Space. In: Shen, D.G., et al., Eds., Medical Image Computing and Computer Assisted Intervention—MICCAI 2019, Springer International Publishing, 695-703. https://doi.org/10.1007/978-3-030-32251-9_76 |
| [9] | Oksuz, I., Clough, J.R., Ruijsink, B., Puyol-Anton, E. and Schnabel, J.A. (2020) Deep Learning-Based Detection and Correction of Cardiac MR Motion Artefacts during Reconstruction for High-Quality Segmentation. IEEE Transactions on Medical Imaging, 39, 4001-4010. |
| [10] | Ke, J., Liu, K., Sun, Y., Xue, Y., Huang, J., Lu, Y., et al. (2023) Artifact Detection and Restoration in Histology Images with Stain-Style and Structural Preservation. IEEE Transactions on Medical Imaging, 42, 3487-3500. https://doi.org/10.1109/tmi.2023.3288940 |
| [11] | Gu, X., Tian, Y., Li, C., Wei, Y. and Li, D. (2024) Improved Se-Resnet Acoustic-Vibration Fusion for Rolling Bearing Composite Fault Diagnosis. Applied Sciences, 14, Article No. 2182. https://doi.org/10.3390/app14052182 |
| [12] | Chen, X., Xing, X., Zhang, Y., Liu, R., Li, L., Zhang, R., et al. (2024) MSCR-FuResNet: A Three-Residual Network Fusion Model Based on Multi-Scale Feature Extraction and Enhanced Channel Spatial Features for Close-Range Apple Leaf Diseases Classification under Optimal Conditions. Horticulturae, 10, Article No. 953. https://doi.org/10.3390/horticulturae10090953 |
| [13] | He, K., Zhang, X., Ren, S. and Sun, J. (2016) Deep Residual Learning for Image Recognition. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, 27-30 June 2016, 770-778. https://doi.org/10.1109/cvpr.2016.90 |
| [14] | Khodadadi Shoushtari, F., Sina, S. and Dehkordi, A.N.V. (2022) Automatic Segmentation of Glioblastoma Multiform Brain Tumor in MRI Images: Using Deeplabv3+ with Pre-Trained Resnet18 Weights. Physica Medica, 100, 51-63. https://doi.org/10.1016/j.ejmp.2022.06.007 |
| [15] | Chen, Y. (2023) Application of Resnet18-Unet in Separating Tumors from Brain MRI Images. Journal of Physics: Conference Series, 2580, Article ID: 012057. https://doi.org/10.1088/1742-6596/2580/1/012057 |
| [16] | Simonyan, K. and Zisserman, A. (2014) Very Deep Convolutional Networks for Large-Scale Image Recognition. arXiv: 1409.1556. |
| [17] | Hu, J., Shen, L., Albanie, S., Sun, G. and Wu, E. (2020) Squeeze-and-Excitation Networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 42, 2011-2023. https://doi.org/10.1109/tpami.2019.2913372 |
| [18] | Huang, G., Liu, Z., Van Der Maaten, L. and Weinberger, K.Q. (2017) Densely Connected Convolutional Networks. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, 21-26 July 2017, 2261-2269. https://doi.org/10.1109/cvpr.2017.243 |
| [19] | Chen, S., Duan, J., Zhang, N., Qi, M., Li, J., Wang, H., et al. (2023) Msa-YOLOv5: Multi-Scale Attention-Based YOLOv5 for Automatic Detection of Acute Ischemic Stroke from Multi-Modality MRI Images. Computers in Biology and Medicine, 165, Article ID: 107471. https://doi.org/10.1016/j.compbiomed.2023.107471 |
