《计算机应用》唯一官方网站 ›› 2025, Vol. 45 ›› Issue (7): 2378-2386.DOI: 10.11772/j.issn.1001-9081.2024070977
• 多媒体计算与计算机仿真 • 上一篇
收稿日期:
2024-07-09
修回日期:
2024-11-19
接受日期:
2024-11-19
发布日期:
2025-07-10
出版日期:
2025-07-10
通讯作者:
柏诗淇
作者简介:
陶永鹏(1981—),男,辽宁大连人,副教授,硕士,CCF会员,主要研究方向:医学图像处理、医疗大数据处理基金资助:
Yongpeng TAO, Shiqi BAI(), Zhengwen ZHOU
Received:
2024-07-09
Revised:
2024-11-19
Accepted:
2024-11-19
Online:
2025-07-10
Published:
2025-07-10
Contact:
Shiqi BAI
About author:
TAO Yongpeng, born in 1981, M. S., associate professor. His research interests include medical image processing, medical big data processing.Supported by:
摘要:
脑胶质瘤在磁共振成像(MRI)图像中的形状大小变化大、边界模糊且组织结构复杂,这些特点导致了脑肿瘤分割任务的挑战性,通常这种任务需要具备深厚专业知识的研究人员设计复杂定制的网络模型才能完成。这一过程不仅耗时,而且需要大量的人力资源。为了简化网络设计流程并自动获取最优的网络结构,提出一种基于卷积和Transformer神经网络架构搜索的脑胶质瘤多组织分割网络(NASCT-Net),以在构建用于多模态MRI脑肿瘤分割的网络架构的过程中,提高分割的精确度。首先,将神经架构搜索(NAS)技术应用于编码器的构建,形成可堆叠的NAS编解码模块,以自动优化适用于脑胶质瘤精准分割的网络架构;其次,在编码器底层集成基于Transformer的特征编码模块,以增强对肿瘤各组之间的相对位置和全局信息的表征能力;最后,通过构建体积加权Dice损失函数(VWDiceLoss),解决前景与背景的不平衡问题。在BraTS2019脑肿瘤数据集上与Swin-Unet等方法进行比较的实验结果表明,NASCT-Net的平均Dice相似系数(DSC)提高了0.009,同时平均Hausdorff距离(HD)降低了1.831 mm,验证了NASCT-Net在提高脑肿瘤多组织分割精度方面的有效性。
中图分类号:
陶永鹏, 柏诗淇, 周正文. 基于卷积和Transformer神经网络架构搜索的脑胶质瘤多组织分割网络[J]. 计算机应用, 2025, 45(7): 2378-2386.
Yongpeng TAO, Shiqi BAI, Zhengwen ZHOU. Neural architecture search for multi-tissue segmentation using convolutional and transformer-based networks in glioma segmentation[J]. Journal of Computer Applications, 2025, 45(7): 2378-2386.
网络结构 | 模块名称 | 输入特征尺寸 | 包含操作 | 输出特征尺寸 |
---|---|---|---|---|
输入 | 输入层 | Resize操作,Concat操作 | ||
卷积层 | ||||
编码器 | NAS-ENC模块 | |||
Transformer模块 | ||||
解码器 | 跳跃连接层 | Concat操作 | ||
NAS-DEC模块 | ||||
输出 | 输出层 |
表1 NASCT-Net各个模块输入输出的维度
Tab. 1 Input and output dimensions of each module in NASCT-Net
网络结构 | 模块名称 | 输入特征尺寸 | 包含操作 | 输出特征尺寸 |
---|---|---|---|---|
输入 | 输入层 | Resize操作,Concat操作 | ||
卷积层 | ||||
编码器 | NAS-ENC模块 | |||
Transformer模块 | ||||
解码器 | 跳跃连接层 | Concat操作 | ||
NAS-DEC模块 | ||||
输出 | 输出层 |
方法 | TC | ET | WT | |||
---|---|---|---|---|---|---|
Dice↑ | HD95/mm↓ | Dice↑ | HD95/mm↓ | Dice↑ | HD95/mm↓ | |
3D U-Net[ | 0.741 | 24.151 | 0.722 | 26.414 | 0.760 | 19.165 |
NAS-Unet[ | 0.754 | 22.237 | 0.740 | 23.986 | 0.808 | 14.327 |
Attention U-Net[ | 0.760 | 21.813 | 0.743 | 24.181 | 0.812 | 15.232 |
Swin-Unet[ | 0.784 | 16.282 | 0.772 | 18.452 | 0.828 | 12.787 |
GCN[ | 0.781 | 17.145 | 0.752 | 20.883 | 0.831 | 12.150 |
NASCT-Net | 0.792 | 15.486 | 0.786 | 17.325 | 0.839 | 10.502 |
表2 在BraTS2019数据集上NASCT-Net与其他方法的HGG分割定量评估结果
Tab. 2 Quantitative evaluation results of HGG segmentation of NASCT-Net and other methods on BraTS2019 dataset
方法 | TC | ET | WT | |||
---|---|---|---|---|---|---|
Dice↑ | HD95/mm↓ | Dice↑ | HD95/mm↓ | Dice↑ | HD95/mm↓ | |
3D U-Net[ | 0.741 | 24.151 | 0.722 | 26.414 | 0.760 | 19.165 |
NAS-Unet[ | 0.754 | 22.237 | 0.740 | 23.986 | 0.808 | 14.327 |
Attention U-Net[ | 0.760 | 21.813 | 0.743 | 24.181 | 0.812 | 15.232 |
Swin-Unet[ | 0.784 | 16.282 | 0.772 | 18.452 | 0.828 | 12.787 |
GCN[ | 0.781 | 17.145 | 0.752 | 20.883 | 0.831 | 12.150 |
NASCT-Net | 0.792 | 15.486 | 0.786 | 17.325 | 0.839 | 10.502 |
方法 | TC | ET | WT | |||
---|---|---|---|---|---|---|
Dice↑ | HD95/mm↓ | Dice↑ | HD95/mm↓ | Dice↑ | HD95/mm↓ | |
3D U-Net[ | 0.693 | 28.732 | 0.726 | 23.975 | 0.753 | 20.215 |
NAS-Unet[ | 0.716 | 26.157 | 0.747 | 24.062 | 0.814 | 16.023 |
Attention U-Net[ | 0.722 | 25.311 | 0.752 | 22.757 | 0.806 | 16.490 |
Swin-Unet[ | 0.746 | 19.175 | 0.787 | 16.674 | 0.818 | 15.462 |
GCN[ | 0.783 | 21.151 | 0.762 | 20.356 | 0.820 | 14.837 |
NASCT-Net | 0.759 | 18.656 | 0.782 | 17.032 | 0.825 | 14.085 |
表3 在BraTS2019数据集上NASCT-Net与其他方法的LGG分割定量评估结果
Tab. 3 Quantitative evaluation results of LGG segmentation of NASCT-Net and other methods on BraTS2019 dataset
方法 | TC | ET | WT | |||
---|---|---|---|---|---|---|
Dice↑ | HD95/mm↓ | Dice↑ | HD95/mm↓ | Dice↑ | HD95/mm↓ | |
3D U-Net[ | 0.693 | 28.732 | 0.726 | 23.975 | 0.753 | 20.215 |
NAS-Unet[ | 0.716 | 26.157 | 0.747 | 24.062 | 0.814 | 16.023 |
Attention U-Net[ | 0.722 | 25.311 | 0.752 | 22.757 | 0.806 | 16.490 |
Swin-Unet[ | 0.746 | 19.175 | 0.787 | 16.674 | 0.818 | 15.462 |
GCN[ | 0.783 | 21.151 | 0.762 | 20.356 | 0.820 | 14.837 |
NASCT-Net | 0.759 | 18.656 | 0.782 | 17.032 | 0.825 | 14.085 |
方法 | TC | ET | WT | |||
---|---|---|---|---|---|---|
Dice↑ | HD95/mm↓ | Dice↑ | HD95/mm↓ | Dice↑ | HD95/mm↓ | |
3D U-Net[ | 0.710 | 26.432 | 0.731 | 25.015 | 0.762 | 19.215 |
NAS-Unet[ | 0.724 | 25.757 | 0.751 | 23.267 | 0.812 | 15.023 |
Attention U-Net[ | 0.731 | 23.014 | 0.763 | 21.032 | 0.818 | 15.615 |
Swin-Unet[ | 0.752 | 19.730 | 0.797 | 18.979 | 0.824 | 14.962 |
GCN[ | 0.748 | 22.878 | 0.782 | 20.137 | 0.827 | 14.533 |
NASCT-Net | 0.765 | 19.056 | 0.812 | 18.017 | 0.836 | 13.757 |
表4 BraTS2021数据集上的脑肿瘤分割定量评估结果
Tab. 4 Quantitative evaluation results of brain tumor segmentation on BraTS2021 dataset
方法 | TC | ET | WT | |||
---|---|---|---|---|---|---|
Dice↑ | HD95/mm↓ | Dice↑ | HD95/mm↓ | Dice↑ | HD95/mm↓ | |
3D U-Net[ | 0.710 | 26.432 | 0.731 | 25.015 | 0.762 | 19.215 |
NAS-Unet[ | 0.724 | 25.757 | 0.751 | 23.267 | 0.812 | 15.023 |
Attention U-Net[ | 0.731 | 23.014 | 0.763 | 21.032 | 0.818 | 15.615 |
Swin-Unet[ | 0.752 | 19.730 | 0.797 | 18.979 | 0.824 | 14.962 |
GCN[ | 0.748 | 22.878 | 0.782 | 20.137 | 0.827 | 14.533 |
NASCT-Net | 0.765 | 19.056 | 0.812 | 18.017 | 0.836 | 13.757 |
模型 | CNN-Transformer结构 | NAS | VWDice损失函数 | Dice | HD95/mm | 参数量/106 |
---|---|---|---|---|---|---|
Baseline | — | — | — | 0.724 | 25.917 | 13.502 |
模型1 | √ | — | — | 0.760 | 20.181 | 16.343 |
模型2 | — | √ | — | 0.756 | 21.042 | 7.140 |
模型3 | — | — | √ | 0.783 | 14.899 | 13.502 |
模型4 | √ | √ | — | 0.794 | 14.026 | 15.065 |
模型5 | √ | — | √ | 0.820 | 12.431 | 16.343 |
模型6 | — | √ | √ | 0.812 | 13.745 | 13.502 |
NASCT-Net | √ | √ | √ | 0.832 | 12.293 | 15.506 |
表5 消融实验的定量评估结果
Tab. 5 Quantitative evaluation results of ablation experiments
模型 | CNN-Transformer结构 | NAS | VWDice损失函数 | Dice | HD95/mm | 参数量/106 |
---|---|---|---|---|---|---|
Baseline | — | — | — | 0.724 | 25.917 | 13.502 |
模型1 | √ | — | — | 0.760 | 20.181 | 16.343 |
模型2 | — | √ | — | 0.756 | 21.042 | 7.140 |
模型3 | — | — | √ | 0.783 | 14.899 | 13.502 |
模型4 | √ | √ | — | 0.794 | 14.026 | 15.065 |
模型5 | √ | — | √ | 0.820 | 12.431 | 16.343 |
模型6 | — | √ | √ | 0.812 | 13.745 | 13.502 |
NASCT-Net | √ | √ | √ | 0.832 | 12.293 | 15.506 |
[1] | SUH J H, KOTECHAR, CHAO S T, et al. Current approaches to the management of brain metastases [J]. Nature Reviews Clinical Oncology, 2020, 17(5): 279-299. |
[2] | KARSCHNIA P, LE RHUN E, VOGELBAUM M A, et al. The evolving role of neurosurgery for central nervous system metastases in the era of personalized cancer therapy [J]. European Journal of Cancer, 2021, 156(7): 93-108. |
[3] | FECCI P E, CHAMPION C D, HOJ J, et al. The evolving modern management of brain metastasis [J]. Clinical Cancer Research, 2019, 25(22): 6570-6580. |
[4] | AOYAMA H, SHIRATO H, TAGO M, et al. Stereotactic radiosurgery plus whole-brain radiation therapy vs stereotactic radiosurgery alone for treatment of brain metastases: a randomized controlled trial [J]. JAMA, 2006, 295(21): 2483-2491. |
[5] | MÜLLER-RIEMENSCHNEIDER F, BOCKELBRINK A, ERNST I, et al. Stereotactic radiosurgery for the treatment of brain metastases [J]. Radiotherapy and Oncology, 2009, 91(1): 67-74. |
[6] | TSAO M N, RADES D, WIRTH A, et al. Radiotherapeutic and surgical management for newly diagnosed brain metastasis (es): an American Society for Radiation Oncology evidence-based guideline [J]. Practical Radiation Oncology, 2012, 2(3): 210-225. |
[7] | CHAO S T, DE SALLES A, HAYASHI M, et al. Stereotactic radiosurgery in the management of limited (1-4) brain metastases: systematic review and international stereotactic radiosurgery society practice guideline [J]. Neurosurgery, 2018, 83(3): 345-353. |
[8] | SAHGAL A, AOYAMA H, KOCHER M, et al. Phase 3 trials of stereotactic radiosurgery with or without whole-brain radiation therapy for 1 to 4 brain metastases: individual patient data meta-analysis [J]. International Journal of Radiation Oncology, Biology, Physics, 2015, 91(4): 710-717. |
[9] | DOLECEK T A, PROPP J M, STROUP N E, et al. CBTRUS statistical report: primary brain and central nervous system tumors diagnosed in the United States in 2005-2009 [J]. Neuro-Oncology, 2012, 14(S5): v1-v49. |
[10] | 田捷,杨鑫,秦承虎,等.光学分子影像技术及其应用[M].北京:科学出版社,2010: 70-106. |
TIAN J, YANG X, QIN C H, et al. Optical molecular imaging technology and its application [M]. Beijing: Science Press, 2010: 70-106. | |
[11] | McGIRT M J, WOODWORTH G F, COON A L, et al. Independent predictors of morbidity after image-guided stereotactic brain biopsy: a risk assessment of 270 cases [J]. Journal of Neurosurgery, 2005, 102(5): 897-901. |
[12] | GRIVALSKY S, TAMAJKA M, BENESOVA W. Segmentation of gliomas in magnetic resonance images using recurrent neural networks [C]// Proceedings of the 42nd International Conference on Telecommunications and Signal Processing. Piscataway: IEEE, 2019: 539-542. |
[13] | GHASSEMI N, SHOEIBI A, ROUHANI M. Deep neural network with generative adversarial networkspre-training for brain tumor classification based on MR images [J]. Biomedical Signal Processing and Control, 2020, 57: No.101678. |
[14] | BALAMURUGAN T, GNANAMANOHARAN E. Brain tumor segmentation and classification using hybrid deep CNN with LuNetClassifier [J]. Neural Computing and Applications, 2023, 35(6): 4739-4753. |
[15] | FARAJZADEH N, SADEGHZADEH N, HASHEMZADEH M. Brain tumor segmentation and classification on MRI via deep hybrid representation learning [J]. Expert Systems with Applications, 2023, 224: No.119963. |
[16] | ANDREWS D W, SCOTT C B, SPERDUTO P W, et al. Whole brain radiation therapy with or without stereotactic radiosurgery boost for patients with one to three brain metastases: phase III results of the RTOG 9508 randomised trial [J]. The Lancet, 2004, 363(9422): 1665-1672. |
[17] | ZHANG Z, YANG J, HO A, et al. A predictive model for distinguishing radiation necrosis from tumor progression after Gamma knife radiosurgery based on radiomic features from MR images [J]. European Radiology, 2018, 28(6): 2255-2263. |
[18] | KEATS L, SITZER D, BALICE G. The relationship between social cognition, neurocognition, and symptomatology on social skill capacity among inpatients with schizophrenia-spectrum disorders [J]. Archives of Clinical Neuropsychology, 2015, 30(6): 501-502. |
[19] | LAMBIN P, RIOS-VELAZQUEZ E, LEIJENAAR R, et al. Radiomics: extracting more information from medical images using advanced feature analysis [J]. European Journal of Cancer, 2012, 48(4): 441-446. |
[20] | VAN VELDEN F H P, KRAMER G M, FRINGS V, et al. Repeatability of radiomic features in non-small-cell lung cancer [(18)F]FDG-PET/CT studies: impact of reconstruction and delineation [J]. Molecular Imaging and Biology, 2016, 18(5): 788-795. |
[21] | GALIMZIANOVA A, PERNUŠ F, LIKAR B, et al. Stratified mixture modeling for segmentation of white-matter lesions in brain MR images [J]. NeuroImage, 2015, 124(Pt A): 1031-1043. |
[22] | WEISS N, RUECKERT D, RAO A. Multiple sclerosis lesion segmentation using dictionary learning and sparse coding [C]// Proceedings of the 2013 International Conference on Medical Image Computing and Computer-Assisted Intervention, LNCS 8149. Berlin: Springer, 2013: 735-742. |
[23] | KARIMAGHALOO Z, RIVAZ H, ARNOLD D L, et al. Adaptive voxel, texture and temporal conditional random fields for detection of gad-enhancing multiple sclerosis lesions in brain MRI [C]// Proceedings of the 2013 International Conference on Medical Image Computing and Computer-Assisted Intervention, LNCS 8151. Berlin: Springer, 2013: 543-550. |
[24] | RONNEBERGER O, FISCHER P, BROX T. U-net: convolutional networks for biomedical image segmentation [C]// Proceedings of the 2015 International Conference on Medical Image Computing and Computer-Assisted Intervention, LNCS 9351. Cham: Springer, 2015: 234-241. |
[25] | OKTAY O, SCHLEMPER J, LE FOLGOCL L, et al. Attention U-Net: learning where to look for the pancreas [EB/OL]. [2024-04-20]. . |
[26] | 朱婷,王瑜,肖洪兵,等.基于多通路CNN的多模态MRI神经胶质瘤分割[J].计算机应用与软件,2018, 35(4): 220-226. |
ZHU T, WANG Y, XIAO H B, et al. Multi-modality MRI gliomas segmentation based on multi-channel CNN [J]. Computer Applications and Software, 2018, 35(4): 220-226. | |
[27] | 李歆,王雪真,洪金省,等.基于图卷积网络的脑胶质瘤核磁共振图像分割[J].计算机系统应用,2024, 33(8): 231-239. |
LI X, WANG X Z, HONG J S, et al. Magnetic resonance image segmentation of gliomas based on graph convolution network [J]. Computer Systems and Applications, 2024, 33(8): 231-239. | |
[28] | CHEN J, LU Y, YU Q, et al. TransUnet: Transformers make strong encoders for medical image segmentation [EB/OL]. [2024-02-08]. . |
[29] | DOSOVITSKIY A, BEYER L, KOLESNIKOV A, et al. An image is worth 16x16 words: Transformers for image recognition at scale [EB/OL]. [2024-06-03]. . |
[30] | LIN J, LIN J, LU C, et al. CKD-TransBTS: clinical knowledge-driven hybrid transformer with modality-correlated cross-attention for brain tumor segmentation [J]. IEEE Transactions on Medical Imaging, 2023, 42(8): 2451-2461. |
[31] | CAO H, WANG Y, CHEN J. Swin-Unet: Unet-like pure transformer for medical image segmentation [C]// Proceedings of the 2022 European Conference on Computer Vision Workshops, LNCS 13803. Cham: Springer, 2022: 205-218. |
[32] | LIN A, CHEN B, XU J, et al. DS-TransUnet: dual Swin Transformer U-Net for medical image segmentation [J]. IEEE Transactions on Instrumentation and Measurement, 2022, 71: No.4005615. |
[33] | SALEHIN I, ISLAM M S, SAHA P, et al. AutoML: a systematic review on automated machine learning with neural architecture search [J]. Journal of Information and Intelligence, 2024, 2(1): 52-81. |
[34] | MA B, ZHANG J, XIA Y, et al. VNAS: variational neural architecture search [J]. International Journal of Computer Vision, 2024, 132(9): 3689-3713. |
[35] | LIU Y, SUN Y, XUE B, et al. A survey on evolutionary neural architecture search [J]. IEEE Transactions on Neural Networks and Learning Systems, 2023, 34(2): 550-570. |
[36] | WENG Y, ZHOU T, LI Y, et al. NAS-Unet: neural architecture search for medical image segmentation [J]. IEEE Access, 2019, 7: 44247-44257. |
[37] | WEI J, FAN Z. Genetic U-Net: automatically designing lightweight U-shaped CNN architectures using the genetic algorithm for retinal vessel segmentation [EB/OL]. [2024-11-02]. . |
[38] | YAN X, JIANG W, SHI Y, et al. MS-NAS: multi-scale neural architecture search for medical image segmentation [C]// Proceedings of the 2020 International Conference on Medical Image and Computing Computer-Assisted Intervention, LNCS 12261. Cham: Springer, 2020: 388-397. |
[39] | LIU X, YAO C, CHEN H, et al. BTSC-TNAS: a neural architecture search-based transformer for brain tumor segmentation and classification [J]. Computerized Medical Imaging and Graphics, 2023, 110: No.102307. |
[40] | XU S, QUAN H. ECT-NAS: searching efficient CNN-Transformers architecture for medical image segmentation [C]// Proceedings of the 2021 IEEE International Conference on Bioinformatics and Biomedicine. Piscataway: IEEE, 2021: 1601-1604. |
[41] | ÇIÇEK Ö, ABDULKADIR A, LIENKAMP S S, et al. 3D U-Net: learning dense volumetric segmentation from sparse annotation [C]// Proceedings of the 2016 International Conference on Medical Image Computing and Computer-Assisted Intervention, LNCS 9901. Cham: Springer, 2016: 424-432. |
[1] | 张英俊, 闫薇薇, 谢斌红, 张睿, 陆望东. 梯度区分与特征范数驱动的开放世界目标检测[J]. 《计算机应用》唯一官方网站, 2025, 45(7): 2203-2210. |
[2] | 吴宗航, 张东, 李冠宇. 基于联合自监督学习的多模态融合推荐算法[J]. 《计算机应用》唯一官方网站, 2025, 45(6): 1858-1868. |
[3] | 龙雨菲, 牟宇辰, 刘晔. 基于张量化图卷积网络和对比学习的多源数据表示学习模型[J]. 《计算机应用》唯一官方网站, 2025, 45(5): 1372-1378. |
[4] | 许鹏程, 何磊, 李川, 钱炜祺, 赵暾. 基于Transformer的深度符号回归方法[J]. 《计算机应用》唯一官方网站, 2025, 45(5): 1455-1463. |
[5] | 王丹, 张文豪, 彭丽娟. 基于深度学习的智能反射面辅助通信系统信道估计[J]. 《计算机应用》唯一官方网站, 2025, 45(5): 1613-1618. |
[6] | 陈凯, 叶海良, 曹飞龙. 基于局部-全局交互与结构Transformer的点云分类算法[J]. 《计算机应用》唯一官方网站, 2025, 45(5): 1671-1676. |
[7] | 陈鹏宇, 聂秀山, 李南君, 李拓. 基于时空解耦和区域鲁棒性增强的半监督视频目标分割方法[J]. 《计算机应用》唯一官方网站, 2025, 45(5): 1379-1386. |
[8] | 李慧, 贾炳志, 王晨曦, 董子宇, 李纪龙, 仲兆满, 陈艳艳. 基于Swin Transformer的生成对抗网络水下图像增强模型[J]. 《计算机应用》唯一官方网站, 2025, 45(5): 1439-1446. |
[9] | 耿海军, 董赟, 胡治国, 池浩田, 杨静, 尹霞. 基于Attention-1DCNN-CE的加密流量分类方法[J]. 《计算机应用》唯一官方网站, 2025, 45(3): 872-882. |
[10] | 袁宝华, 陈佳璐, 王欢. 融合多尺度语义和双分支并行的医学图像分割网络[J]. 《计算机应用》唯一官方网站, 2025, 45(3): 988-995. |
[11] | 王地欣, 王佳昊, 李敏, 陈浩, 胡光耀, 龚宇. 面向水声通信网络的异常攻击检测[J]. 《计算机应用》唯一官方网站, 2025, 45(2): 526-533. |
[12] | 张翰林, 王俊陆, 宋宝燕. 融合衍生特征的时间序列事件分类方法[J]. 《计算机应用》唯一官方网站, 2025, 45(2): 428-435. |
[13] | 王雅伦, 张仰森, 朱思文. 面向知识推理的位置编码标题生成模型[J]. 《计算机应用》唯一官方网站, 2025, 45(2): 345-353. |
[14] | 徐欣然, 张绍兵, 成苗, 张洋, 曾尚. 基于多路层次化混合专家模型的轴承故障诊断方法[J]. 《计算机应用》唯一官方网站, 2025, 45(1): 59-68. |
[15] | 梁杰涛, 罗兵, 付兰慧, 常青玲, 李楠楠, 易宁波, 冯其, 何鑫, 邓辅秦. 基于坐标几何采样的点云配准方法[J]. 《计算机应用》唯一官方网站, 2025, 45(1): 214-222. |
阅读次数 | ||||||
全文 |
|
|||||
摘要 |
|
|||||