Point features, as the basis of lines, surfaces, and bodies, are commonly used in medical image registration. To obtain an elegant spatial transformation of extracted feature points, many point set matching algorithms...Point features, as the basis of lines, surfaces, and bodies, are commonly used in medical image registration. To obtain an elegant spatial transformation of extracted feature points, many point set matching algorithms(PMs) have been developed to match two point sets by optimizing multifarious distance functions. There are ample reviews related to medical image registration and PMs which summarize their basic principles and main algorithms separately. However,to data, detailed summary of PMs used in medical image registration in different clinical environments has not been published. In this paper, we provide a comprehensive review of the existing key techniques of the PMs applied to medical image registration according to the basic principles and clinical applications. As the core technique of the PMs, geometric transformation models are elaborated in this paper, demonstrating the mechanism of point set registration. We also focus on the clinical applications of the PMs and propose a practical classification method according to their applications in different clinical surgeries. The aim of this paper is to provide a summary of pointfeaturebased methods used in medical image registration and to guide doctors or researchers interested in this field to choose appropriate techniques in their research.展开更多
Currently the voxel based registration methods have been used widely such as the well known mutual information (MI). Although the accuracy of their results is plausible, the registration procedure is slow. This paper ...Currently the voxel based registration methods have been used widely such as the well known mutual information (MI). Although the accuracy of their results is plausible, the registration procedure is slow. This paper proposed some methods to rigid registration based on mutual information, aiming for an acceleration of the registration process without significantly loss of precision in the final results. The efficiency of these methods is examined by registration of CT MR and PET MR. Experimental results show that the speedup is effective and efficient. By using the fast methods, the registration of 3 D medical image could also be implemented on PC rapidly.展开更多
In order to improve the registration accuracy of brain magnetic resonance images(MRI),some deep learning registration methods use segmentation images for training model.How-ever,the segmentation values are constant fo...In order to improve the registration accuracy of brain magnetic resonance images(MRI),some deep learning registration methods use segmentation images for training model.How-ever,the segmentation values are constant for each label,which leads to the gradient variation con-centrating on the boundary.Thus,the dense deformation field(DDF)is gathered on the boundary and there even appears folding phenomenon.In order to fully leverage the label information,the morphological opening and closing information maps are introduced to enlarge the non-zero gradi-ent regions and improve the accuracy of DDF estimation.The opening information maps supervise the registration model to focus on smaller,narrow brain regions.The closing information maps supervise the registration model to pay more attention to the complex boundary region.Then,opening and closing morphology networks(OC_Net)are designed to automatically generate open-ing and closing information maps to realize the end-to-end training process.Finally,a new registra-tion architecture,VM_(seg+oc),is proposed by combining OC_Net and VoxelMorph.Experimental results show that the registration accuracy of VM_(seg+oc) is significantly improved on LPBA40 and OASIS1 datasets.Especially,VM_(seg+oc) can well improve registration accuracy in smaller brain regions and narrow regions.展开更多
In image-guided radiation therapy, extracting features from medical point cloud is the key technique for multimodality registration. This novel framework, denoted Control Point Net (CPN), provides an alternative to th...In image-guided radiation therapy, extracting features from medical point cloud is the key technique for multimodality registration. This novel framework, denoted Control Point Net (CPN), provides an alternative to the common applications of manually designed keypoint descriptors for coarse point cloud registration. The CPN directly consumes a point cloud, divides it into equally spaced 3D voxels and transforms the points within each voxel into a unified feature representation through voxel feature encoding (VFE) layer. Then all volumetric representations are aggregated by Weighted Extraction Layer which selectively extracts features and synthesize into global descriptors and coordinates of control points. Utilizing global descriptors instead of local features allows the available geometrical data to be better exploited to improve the robustness and precision. Specifically, CPN unifies feature extraction and clustering into a single network, omitting time-consuming feature matching procedure. The algorithm is tested on point cloud datasets generated from CT images. Experiments and comparisons with the state-of-the-art descriptors demonstrate that CPN is highly discriminative, efficient, and robust to noise and density changes.展开更多
A new implementation of the image registration algorithm based on the mutual information is presented for the case of medical images. The registration is achieved if the maximum of the mutual information is attained. ...A new implementation of the image registration algorithm based on the mutual information is presented for the case of medical images. The registration is achieved if the maximum of the mutual information is attained. In this maximization process optimal values of five parameters of an affine transformation are searched.展开更多
A mutual information based 3D non-rigid registration approach was proposed for the registration of deformable CT/MR body abdomen images. The Parzen Windows Density Estimation (PWDE) method is adopted to calculate the ...A mutual information based 3D non-rigid registration approach was proposed for the registration of deformable CT/MR body abdomen images. The Parzen Windows Density Estimation (PWDE) method is adopted to calculate the mutual information between the two modals of CT and MRI abdomen images. By maximizing MI between the CT and MR volume images, the overlapping part of them reaches the biggest, which means that the two body images of CT and MR matches best to each other. Visible Human Project (VHP) Male abdomen CT and MRI Data are used as experimental data sets. The experimental results indicate that this approach of non-rigid 3D registration of CT/MR body abdominal images can be achieved effectively and automatically, without any prior processing procedures such as segmentation and feature extraction, but has a main drawback of very long computation time.展开更多
Mutual information has currently been one of the most intensivelyresearched measures. It has been proven to be accurate and effective registrationmeasure. Despite the general promising results, mutual information some...Mutual information has currently been one of the most intensivelyresearched measures. It has been proven to be accurate and effective registrationmeasure. Despite the general promising results, mutual information sometimes mightlead to misregistration because of neglecting spatial information and treating intensityvariations with undue sensitivity. In this paper, an extension of mutual informationframework was proposed in which higher-order spatial information regarding imagestructures was incorporated into the registration processing of PET and MR. Thesecond-order estimate of mutual information algorithm was applied to the registrationof seven patients. Evaluation from Vanderbilt University and our visual inspectionshowed that sub-voxel accuracy and robust results were achieved in all cases withsecond-order mutual information as the similarity measure and with Powell's multi-dimensional direction set method as optimization strategy.展开更多
Many deep learning-based registration methods rely on a single-stream encoder-decoder network for computing deformation fields between 3D volumes.However,these methods often lack constraint information and overlook se...Many deep learning-based registration methods rely on a single-stream encoder-decoder network for computing deformation fields between 3D volumes.However,these methods often lack constraint information and overlook semantic consistency,limiting their performance.To address these issues,we present a novel approach for medical image registration called theDual-VoxelMorph,featuring a dual-channel cross-constraint network.This innovative network utilizes both intensity and segmentation images,which share identical semantic information and feature representations.Two encoder-decoder structures calculate deformation fields for intensity and segmentation images,as generated by the dual-channel cross-constraint network.This design facilitates bidirectional communication between grayscale and segmentation information,enabling the model to better learn the corresponding grayscale and segmentation details of the same anatomical structures.To ensure semantic and directional consistency,we introduce constraints and apply the cosine similarity function to enhance semantic consistency.Evaluation on four public datasets demonstrates superior performance compared to the baselinemethod,achieving Dice scores of 79.9%,64.5%,69.9%,and 63.5%for OASIS-1,OASIS-3,LPBA40,and ADNI,respectively.展开更多
Significant breakthroughs in medical image registration have been achieved using deep neural networks(DNNs).However,DNN-based end-to-end registration methods often require large quantities of data or adequate annotati...Significant breakthroughs in medical image registration have been achieved using deep neural networks(DNNs).However,DNN-based end-to-end registration methods often require large quantities of data or adequate annotations for training.To leverage the intensity information of abundant unlabeled images,unsupervised registration methods commonly employ intensity-based similarity measures to optimize the network parameters.However,finding a sufficiently robust measure can be challenging for specific registration applications.Weakly supervised registration methods use anatomical labels to estimate the deformation between images.High-level structural information in label images is more reliable and practical for estimating the voxel correspondence of anatomic regions of interest between images,whereas label images are extremely difficult to collect.In this paper,we propose a two-stage semi-supervised learning framework for medical image registration,which consists of unsupervised and weakly supervised registration networks.The proposed semi-supervised learning framework is trained with intensity information from available images,label information from a relatively small number of labeled images and pseudo-label information from unlabeled images.Experimental results on two datasets(cardiac and abdominal images)demonstrate the efficacy and efficiency of this method in intra-and inter-modality medical image registrations,as well as its superior performance when a vast amount of unlabeled data and a small set of annotations are available.Our code is publicly available at at https://github.com/jdq818/SeRN.展开更多
Transformers have dominated the field of natural language processing and have recently made an impact in the area of computer vision.In the field of medical image analysis,transformers have also been successfully used...Transformers have dominated the field of natural language processing and have recently made an impact in the area of computer vision.In the field of medical image analysis,transformers have also been successfully used in to full-stack clinical applications,including image synthesis/reconstruction,registration,segmentation,detection,and diagnosis.This paper aimed to promote awareness of the applications of transformers in medical image analysis.Specifically,we first provided an overview of the core concepts of the attention mechanism built into transformers and other basic components.Second,we reviewed various transformer architectures tailored for medical image applications and discuss their limitations.Within this review,we investigated key challenges including the use of transformers in different learning paradigms,improving model efficiency,and coupling with other techniques.We hope this review would provide a comprehensive picture of transformers to readers with an interest in medical image analysis.展开更多
In landmark-based image registration, estimating the landmark correspondence plays an important role. In this letter, a novel landmark correspondence estimation technique using mean shift algorithm is proposed. Image ...In landmark-based image registration, estimating the landmark correspondence plays an important role. In this letter, a novel landmark correspondence estimation technique using mean shift algorithm is proposed. Image corner points are detected as landmarks and mean shift iterations are adopted to find the most probable corresponding point positions in two images. Mutual information between intensity of two local regions is computed to eliminate mis-matching points to improve the stability of corresponding estimation correspondence landmarks is exact. The proposed experiments of various mono-modal medical images. Multi-level estimation (MLE) technique is proposed Experiments show that the precision in location of technique is shown to be feasible and rapid in the展开更多
The early implementation of treatment therapies necessitates the swift and precise identification of COVID-19 pneumonia by the analysis of chest CT scans.This study aims to investigate the indispensable need for preci...The early implementation of treatment therapies necessitates the swift and precise identification of COVID-19 pneumonia by the analysis of chest CT scans.This study aims to investigate the indispensable need for precise and interpretable diagnostic tools for improving clinical decision-making for COVID-19 diagnosis.This paper proposes a novel deep learning approach,called Conformer Network,for explainable discrimination of viral pneumonia depending on the lung Region of Infections(ROI)within a single modality radiographic CT scan.Firstly,an efficient U-shaped transformer network is integrated for lung image segmentation.Then,a robust transfer learning technique is introduced to design a robust feature extractor based on pre-trained lightweight Big Transfer(BiT-L)and finetuned on medical data to effectively learn the patterns of infection in the input image.Secondly,this work presents a visual explanation method to guarantee clinical explainability for decisions made by Conformer Network.Experimental evaluation of real-world CT data demonstrated that the diagnostic accuracy of ourmodel outperforms cutting-edge studies with statistical significance.The Conformer Network achieves 97.40% of detection accuracy under cross-validation settings.Our model not only achieves high sensitivity and specificity but also affords visualizations of salient features contributing to each classification decision,enhancing the overall transparency and trustworthiness of our model.The findings provide obvious implications for the ability of our model to empower clinical staff by generating transparent intuitions about the features driving diagnostic decisions.展开更多
经典的UNet网络可用于预测全分辨率空间域的密集位移场,在医学图像配准中取得了巨大成功。但对大变形的三维图像配准,还存在运行时间长、无法有效保持拓扑结构、空间特征易丢失等缺点。为此,提出一种基于交叉注意的大内核多尺度可变形...经典的UNet网络可用于预测全分辨率空间域的密集位移场,在医学图像配准中取得了巨大成功。但对大变形的三维图像配准,还存在运行时间长、无法有效保持拓扑结构、空间特征易丢失等缺点。为此,提出一种基于交叉注意的大内核多尺度可变形医学图像配准网络(large kernel multi-scale deformable medical image registration network based on cross-attention,LK-CAUNet)。在经典UNet模型基础上,通过引入交叉注意力模块,实现高效、多层次的语义特征融合;配备大内核非对称并行卷积,使其具有多尺度特征和对复杂结构的学习能力;通过加入平方和缩放模块,实现拓扑守恒和变换可逆。基于脑部MRI数据集,将LK-CAUNet与18种经典图像配准模型进行了比较,结果表明,LK-CAUNet的配准性能较其他模型有明显提升,其Dice得分较TransMorph配准方法提高了8%,而参数量仅为TransMorph的1/5。展开更多
基金Supported by the National Natural Science Foundation of China(Grant No.61533016)
文摘Point features, as the basis of lines, surfaces, and bodies, are commonly used in medical image registration. To obtain an elegant spatial transformation of extracted feature points, many point set matching algorithms(PMs) have been developed to match two point sets by optimizing multifarious distance functions. There are ample reviews related to medical image registration and PMs which summarize their basic principles and main algorithms separately. However,to data, detailed summary of PMs used in medical image registration in different clinical environments has not been published. In this paper, we provide a comprehensive review of the existing key techniques of the PMs applied to medical image registration according to the basic principles and clinical applications. As the core technique of the PMs, geometric transformation models are elaborated in this paper, demonstrating the mechanism of point set registration. We also focus on the clinical applications of the PMs and propose a practical classification method according to their applications in different clinical surgeries. The aim of this paper is to provide a summary of pointfeaturebased methods used in medical image registration and to guide doctors or researchers interested in this field to choose appropriate techniques in their research.
文摘Currently the voxel based registration methods have been used widely such as the well known mutual information (MI). Although the accuracy of their results is plausible, the registration procedure is slow. This paper proposed some methods to rigid registration based on mutual information, aiming for an acceleration of the registration process without significantly loss of precision in the final results. The efficiency of these methods is examined by registration of CT MR and PET MR. Experimental results show that the speedup is effective and efficient. By using the fast methods, the registration of 3 D medical image could also be implemented on PC rapidly.
基金supported by Shandong Provincial Natural Science Foundation(No.ZR2023MF062)the National Natural Science Foundation of China(No.61771230).
文摘In order to improve the registration accuracy of brain magnetic resonance images(MRI),some deep learning registration methods use segmentation images for training model.How-ever,the segmentation values are constant for each label,which leads to the gradient variation con-centrating on the boundary.Thus,the dense deformation field(DDF)is gathered on the boundary and there even appears folding phenomenon.In order to fully leverage the label information,the morphological opening and closing information maps are introduced to enlarge the non-zero gradi-ent regions and improve the accuracy of DDF estimation.The opening information maps supervise the registration model to focus on smaller,narrow brain regions.The closing information maps supervise the registration model to pay more attention to the complex boundary region.Then,opening and closing morphology networks(OC_Net)are designed to automatically generate open-ing and closing information maps to realize the end-to-end training process.Finally,a new registra-tion architecture,VM_(seg+oc),is proposed by combining OC_Net and VoxelMorph.Experimental results show that the registration accuracy of VM_(seg+oc) is significantly improved on LPBA40 and OASIS1 datasets.Especially,VM_(seg+oc) can well improve registration accuracy in smaller brain regions and narrow regions.
文摘In image-guided radiation therapy, extracting features from medical point cloud is the key technique for multimodality registration. This novel framework, denoted Control Point Net (CPN), provides an alternative to the common applications of manually designed keypoint descriptors for coarse point cloud registration. The CPN directly consumes a point cloud, divides it into equally spaced 3D voxels and transforms the points within each voxel into a unified feature representation through voxel feature encoding (VFE) layer. Then all volumetric representations are aggregated by Weighted Extraction Layer which selectively extracts features and synthesize into global descriptors and coordinates of control points. Utilizing global descriptors instead of local features allows the available geometrical data to be better exploited to improve the robustness and precision. Specifically, CPN unifies feature extraction and clustering into a single network, omitting time-consuming feature matching procedure. The algorithm is tested on point cloud datasets generated from CT images. Experiments and comparisons with the state-of-the-art descriptors demonstrate that CPN is highly discriminative, efficient, and robust to noise and density changes.
文摘A new implementation of the image registration algorithm based on the mutual information is presented for the case of medical images. The registration is achieved if the maximum of the mutual information is attained. In this maximization process optimal values of five parameters of an affine transformation are searched.
基金An international cooperation project between Shanghai Jiaotong U niversity and Hong Kong Polytechnic University
文摘A mutual information based 3D non-rigid registration approach was proposed for the registration of deformable CT/MR body abdomen images. The Parzen Windows Density Estimation (PWDE) method is adopted to calculate the mutual information between the two modals of CT and MRI abdomen images. By maximizing MI between the CT and MR volume images, the overlapping part of them reaches the biggest, which means that the two body images of CT and MR matches best to each other. Visible Human Project (VHP) Male abdomen CT and MRI Data are used as experimental data sets. The experimental results indicate that this approach of non-rigid 3D registration of CT/MR body abdominal images can be achieved effectively and automatically, without any prior processing procedures such as segmentation and feature extraction, but has a main drawback of very long computation time.
基金The images and the standard transformation were provided as part of the project,"Retrospective Im-age Registration Evaluation"(National Institutes of Health,1 R01 CA89323),the principal investigator,J.Michael Fitzpatrick,Vanderbilt Universi
文摘Mutual information has currently been one of the most intensivelyresearched measures. It has been proven to be accurate and effective registrationmeasure. Despite the general promising results, mutual information sometimes mightlead to misregistration because of neglecting spatial information and treating intensityvariations with undue sensitivity. In this paper, an extension of mutual informationframework was proposed in which higher-order spatial information regarding imagestructures was incorporated into the registration processing of PET and MR. Thesecond-order estimate of mutual information algorithm was applied to the registrationof seven patients. Evaluation from Vanderbilt University and our visual inspectionshowed that sub-voxel accuracy and robust results were achieved in all cases withsecond-order mutual information as the similarity measure and with Powell's multi-dimensional direction set method as optimization strategy.
基金National Natural Science Foundation of China(Grant Nos.62171130,62172197,61972093)the Natural Science Foundation of Fujian Province(Grant Nos.2020J01573,2022J01131257,2022J01607)+3 种基金Fujian University Industry University Research Joint Innovation Project(No.2022H6006)in part by the Fund of Cloud Computing and BigData for SmartAgriculture(GrantNo.117-612014063)NationalNatural Science Foundation of China(Grant No.62301160)Nature Science Foundation of Fujian Province(Grant No.2022J01607).
文摘Many deep learning-based registration methods rely on a single-stream encoder-decoder network for computing deformation fields between 3D volumes.However,these methods often lack constraint information and overlook semantic consistency,limiting their performance.To address these issues,we present a novel approach for medical image registration called theDual-VoxelMorph,featuring a dual-channel cross-constraint network.This innovative network utilizes both intensity and segmentation images,which share identical semantic information and feature representations.Two encoder-decoder structures calculate deformation fields for intensity and segmentation images,as generated by the dual-channel cross-constraint network.This design facilitates bidirectional communication between grayscale and segmentation information,enabling the model to better learn the corresponding grayscale and segmentation details of the same anatomical structures.To ensure semantic and directional consistency,we introduce constraints and apply the cosine similarity function to enhance semantic consistency.Evaluation on four public datasets demonstrates superior performance compared to the baselinemethod,achieving Dice scores of 79.9%,64.5%,69.9%,and 63.5%for OASIS-1,OASIS-3,LPBA40,and ADNI,respectively.
文摘Significant breakthroughs in medical image registration have been achieved using deep neural networks(DNNs).However,DNN-based end-to-end registration methods often require large quantities of data or adequate annotations for training.To leverage the intensity information of abundant unlabeled images,unsupervised registration methods commonly employ intensity-based similarity measures to optimize the network parameters.However,finding a sufficiently robust measure can be challenging for specific registration applications.Weakly supervised registration methods use anatomical labels to estimate the deformation between images.High-level structural information in label images is more reliable and practical for estimating the voxel correspondence of anatomic regions of interest between images,whereas label images are extremely difficult to collect.In this paper,we propose a two-stage semi-supervised learning framework for medical image registration,which consists of unsupervised and weakly supervised registration networks.The proposed semi-supervised learning framework is trained with intensity information from available images,label information from a relatively small number of labeled images and pseudo-label information from unlabeled images.Experimental results on two datasets(cardiac and abdominal images)demonstrate the efficacy and efficiency of this method in intra-and inter-modality medical image registrations,as well as its superior performance when a vast amount of unlabeled data and a small set of annotations are available.Our code is publicly available at at https://github.com/jdq818/SeRN.
基金the National Natural Science Foundation of China(Grant No.62106101)the Natural Science Foundation of Jiangsu Province(Grant No.BK20210180).
文摘Transformers have dominated the field of natural language processing and have recently made an impact in the area of computer vision.In the field of medical image analysis,transformers have also been successfully used in to full-stack clinical applications,including image synthesis/reconstruction,registration,segmentation,detection,and diagnosis.This paper aimed to promote awareness of the applications of transformers in medical image analysis.Specifically,we first provided an overview of the core concepts of the attention mechanism built into transformers and other basic components.Second,we reviewed various transformer architectures tailored for medical image applications and discuss their limitations.Within this review,we investigated key challenges including the use of transformers in different learning paradigms,improving model efficiency,and coupling with other techniques.We hope this review would provide a comprehensive picture of transformers to readers with an interest in medical image analysis.
基金supported by the National Natural Science Foundation of China under Grant No.60572101
文摘In landmark-based image registration, estimating the landmark correspondence plays an important role. In this letter, a novel landmark correspondence estimation technique using mean shift algorithm is proposed. Image corner points are detected as landmarks and mean shift iterations are adopted to find the most probable corresponding point positions in two images. Mutual information between intensity of two local regions is computed to eliminate mis-matching points to improve the stability of corresponding estimation correspondence landmarks is exact. The proposed experiments of various mono-modal medical images. Multi-level estimation (MLE) technique is proposed Experiments show that the precision in location of technique is shown to be feasible and rapid in the
基金funded by King Saud University,Riyadh,Saudi Arabia.Researchers Supporting Project Number(RSP2024R167),King Saud University,Riyadh,Saudi Arabia.
文摘The early implementation of treatment therapies necessitates the swift and precise identification of COVID-19 pneumonia by the analysis of chest CT scans.This study aims to investigate the indispensable need for precise and interpretable diagnostic tools for improving clinical decision-making for COVID-19 diagnosis.This paper proposes a novel deep learning approach,called Conformer Network,for explainable discrimination of viral pneumonia depending on the lung Region of Infections(ROI)within a single modality radiographic CT scan.Firstly,an efficient U-shaped transformer network is integrated for lung image segmentation.Then,a robust transfer learning technique is introduced to design a robust feature extractor based on pre-trained lightweight Big Transfer(BiT-L)and finetuned on medical data to effectively learn the patterns of infection in the input image.Secondly,this work presents a visual explanation method to guarantee clinical explainability for decisions made by Conformer Network.Experimental evaluation of real-world CT data demonstrated that the diagnostic accuracy of ourmodel outperforms cutting-edge studies with statistical significance.The Conformer Network achieves 97.40% of detection accuracy under cross-validation settings.Our model not only achieves high sensitivity and specificity but also affords visualizations of salient features contributing to each classification decision,enhancing the overall transparency and trustworthiness of our model.The findings provide obvious implications for the ability of our model to empower clinical staff by generating transparent intuitions about the features driving diagnostic decisions.
文摘经典的UNet网络可用于预测全分辨率空间域的密集位移场,在医学图像配准中取得了巨大成功。但对大变形的三维图像配准,还存在运行时间长、无法有效保持拓扑结构、空间特征易丢失等缺点。为此,提出一种基于交叉注意的大内核多尺度可变形医学图像配准网络(large kernel multi-scale deformable medical image registration network based on cross-attention,LK-CAUNet)。在经典UNet模型基础上,通过引入交叉注意力模块,实现高效、多层次的语义特征融合;配备大内核非对称并行卷积,使其具有多尺度特征和对复杂结构的学习能力;通过加入平方和缩放模块,实现拓扑守恒和变换可逆。基于脑部MRI数据集,将LK-CAUNet与18种经典图像配准模型进行了比较,结果表明,LK-CAUNet的配准性能较其他模型有明显提升,其Dice得分较TransMorph配准方法提高了8%,而参数量仅为TransMorph的1/5。