[1]郭孜文,付 庄夘,张志谊.基于隐式神经网络的超声图像重建方法[J].机械与电子,2025,(11):3-7.
　GUO Ziwen,FU Zhuang,ZHANG Zhiyi.Implicit Neural Representation-based Ultrasound Image Reconstruction[J].Machinery & Electronics,2025,(11):3-7.
点击复制

基于隐式神经网络的超声图像重建方法()
分享到：

《机械与电子》[ISSN:1001-2257/CN:52-1052/TH]

卷:
期数:
2025年11期
页码:
3-7
栏目:
研究与设计
出版日期:
2025-11-24

文章信息/Info

Title:
Implicit Neural Representation-based Ultrasound Image Reconstruction
文章编号:
1001-2257 ( 2025 ) 11-0003-05
作者:
郭孜文; 付 庄夘; 张志谊
上海交通大学机械与动力工程学院,上海 200240
Author(s):
GUO Ziwen ;  FU Zhuang ;  ZHANG Zhiyi
( School of Mechanical Engineering , Shanghai Jiao Tong University , Shanghai 200240 , China )
关键词:
超声成像; 隐式神经网络; 神经辐射场; 体积渲染
Keywords:
ultrasound imaging ;  implicit neural representation ;  neural radiance field ;  volume rendering
分类号:
TP391.4
文献标志码:
A
摘要:
针对传统二维超声难以呈现组织三维结构,且图像质量受探头角度和物理建模不足限制的问题,提出一种结合隐式神经网络与超声成像模型的三维重建方法。通过构建以空间坐标为输入的 MLP 网络,预测介质物理属性,并结合位置编码增强高频细节表达。同时融入能量衰减、反射与散射等物理特性建模,提高新视角图像的真实感与清晰度。实验结果显示,该方法在 SSIM 、 LPIPS 和 PSNR 等指标上优于传统体素插值和 NeRF 方法,显著提升了图像结构还原能力与视觉质量。
Abstract:
To address the challenges of traditional 2D ultrasound in representing 3D tissue structures and its dependence on probe angles and limited physical modeling , this paper proposes a 3D reconstruction method combining implicit neural networks with ultrasound imaging models.By constructing an MLP network with spatial coordinates as input , it predicts physical properties of the medium and enhances high frequency detail expression through positional encoding.The integration of energy attenuation , reflection , and scattering modeling improves the realism and clarity of images from new viewpoints.Experimental results show that this method outperforms traditional voxel interpolation and NeRF in SSIM , LPIPS and PSNR , significantly enhancing structural fidelity and visual quality.

参考文献/References:

[ 1 ] FENSTER A , DOWNEY D B , CARDINAL H N.Three dimensional ultrasound imaging [ J ] .Physics in medicine and biology , 2001 , 46 ( 3 ):67-99.

[ 2 ] DYER C R.Volumetric scene reconstruction from multiple views [ C ] ∥Foundations of Image Understanding , 2001 : 469-489.
[ 3 ] MERCIER L , LANG? T , LINDSETH F , et al.A review of calibration techniques for freehand 3-D ultrasound systems [ J ] .Ultrasound in medicine and biology , 2005 , 31 ( 4 ): 449-471.
[ 4 ] MICHALKIEWICZ M , PONTES J K , JACK D , et al.Implicit surface representations as layers in neural networks [ C ] ∥Proceedings of the IEEE / CVF International Conference on Computer Vision , 2019 : 4743-4752.
[ 5 ] GU F , CHANG H , ZHU W , et al.Implicit graph neural networks [ J ] .Advances in neural information processing systems , 2020 , 33 : 11984-11995.
[ 6 ] MILDENHALL B , SRINIVASAN P P , TANCIK M , et al.NeRF : representing scenes as neural radiance fields for view synthesis [ J ] .Communications of the ACM , 2021 , 65 ( 1 ): 99-106.
[ 7 ] M?LLER T , EVANS A , SCHIED C , et al.Instant neural graphics primitives with a multiresolution hash encoding [ J ] .ACM Transactions on graphics ( TOG ), 2022 , 41 ( 4 ): 1-15.
[ 8 ] PUMAROLA A , CORONA E , PONS-MOLL G , et al. D-nerf : Neural radiance fields for dynamic scenes [ C ] ∥ Proceedings of the IEEE / CVF Conference on Computer Vision and Pattern Recognition , 2021 : 10318-10327.
[ 9 ] 郑太雄,黄帅,李永福,等 . 基于视觉的三维重建关键技术研究综述[ J ] . 自动化学报, 2020 , 46 ( 4 ): 631-652.
[ 10 ] 李吉洋,程乐超,何靖璇,等 . 神经辐射场的研究现状与展望[ J ] . 计算机辅助设计与图形学学报, 2024 , 36( 7 ): 995-1013.
[ 11 ] MOLAEI A , AMINIMEHR A , TAVAKOLI A , et al. Implicit neural representation in medical imaging : a comparative survey [ C ] ∥Proceedings of the IEEE / CVF International Conference on Computer Vision , 2023 : 2381-2391.
[ 12 ] CORONA-FIGUEROA A , FRAWLEY J , BOND TAYLOR S , et al.MedNeRF : medical neural radiance fields for reconstructing 3D-aware CT-projections from a single X-ray [ C ] ∥2022 44th Annual International Conference of the IEEE Engineering in Medicine and Biology Society ( EMBC ), New York : IEEE , 2022 : 3843-3848.
[ 13 ] WANG Y , LONG Y , FAN S H , et al.Neural rendering for stereo 3d reconstruction of deformable tissues in robotic surgery [ C ] ∥International Conference on Medical Image Computing and Computer-assisted Intervention , 2022 : 431-441.
[ 14 ] BARRON J T , MILDENHALL B , TANCIK M , et al. Mip-NeRF : a multiscale representation for anti aliasing neural radiance fields [ C ] ∥Proceedings of the IEEE / CVF International Conference on Computer Vision , 2021 : 5855-5864.
[ 15 ] VERBIN D , HEDMAN P , MILDENHALL B , et al. Ref-NeRF : structured view-dependent appearance for neural radiance fields [ C ] ∥2022 IEEE / CVF Conference on Computer Vision and Pattern Recognition ( CVPR ), 2022 : 5481-5490.
[ 16 ] TANCIK M , SRINIVASAN P , MILDENHALL B , et al. Fourier features let networks learn high frequency functions in low dimensional domains [ J ] .Advances in neural information processing systems , 2020 , 33 : 7537-7547.
[ 17 ] SALEHI M , AHMADI S A , PREVOST R , et al.Patient-specific 3D ultrasound simulation based on convolutional ray-tracing and appearance optimization [ C ] ∥ 18th International Conference of Medical Image Computing and Computer-Assisted Intervention ( MICCAI ), 2015 : 510-518.
[ 18 ] SZABO T L.Diagnostic ultrasound imaging : inside out [ M ] .New York : Academic press , 2013.
[ 19 ] MERCIER L , DEL MAESTRO R F , PETRECCA K , et al.Online database of clinical MR and ultrasound images of brain tumors [ J ] .Medical physics , 2012 , 39( 6 ): 3253-3261.
[ 20 ] ZHANG R , ISOLA P , EFROS A A , et al.The unreasonable effectiveness of deep features as a perceptual metric [ C ] ∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , 2018 : 586-595.

备注/Memo

备注/Memo:
收稿日期: 2025-03-13
基金项目:医工交叉项目( YG2019ZDA17 , ZH2018QNB23 );国家自然科学基金面上项目( 61973210 )
作者简介:郭孜文 ( 2000- ),男,安徽芜湖人,硕士研究生,研究方向为医学图像处理;付 庄 ( 1972- ),男,山东招远人,教授,博士研究生导师,研究方向为机器人及智能控制系统、运动学与动力学等,通信作者, E-mail : zhfu@sjtu.edu.cn ;张志谊 ( 1970- ),男,安徽无为人,博士研究生导师,研究方向为振动主动控制。
更新日期/Last Update: 2025-12-10