Free form deformation and symmetry constraint‐based multimodal brain image registration using generative adversarial nets

Multi‐modal brain image registration has been widely applied to functional localisation,neurosurgery and computational anatomy.The existing registration methods based on the dense deformation fields involve too many parameters,which is not conducive to the exploration of correct spatial correspondence between the float and reference images.Meanwhile,the unidirectional registration may involve the deformation folding,which will result in the change of topology during registration.To address these issues,this work has presented an unsupervised image registration method using the free form deformation(FFD)and the symmetry constraint‐based generative adversarial networks(FSGAN).The FSGAN utilises the principle component analysis network‐based structural representations of the reference and float images as the inputs and uses the generator to learn the FFD model parameters,thereby producing two deformation fields.Meanwhile,the FSGAN uses two discriminators to decide whether the bilateral registration have been realised simultaneously.Besides,the symmetry constraint is utilised to construct the loss function,thereby avoiding the deformation folding.Experiments on BrainWeb,high grade gliomas,IXI and LPBA40 show that compared with state‐of‐the‐art methods,the FSGAN provides superior performance in terms of visual comparisons and such quantitative indexes as dice value,target registration error and computational efficiency.

Application of Opening and Closing Morphology in Deep Learning-Based Brain Image Registration

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.

A Review of Point Feature Based Medical Image Registration

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.

SuperFusion: A Versatile Image Registration and Fusion Network with Semantic Awareness

Image fusion aims to integrate complementary information in source images to synthesize a fused image comprehensively characterizing the imaging scene. However, existing image fusion algorithms are only applicable to strictly aligned source images and cause severe artifacts in the fusion results when input images have slight shifts or deformations. In addition,the fusion results typically only have good visual effect, but neglect the semantic requirements of high-level vision tasks.This study incorporates image registration, image fusion, and semantic requirements of high-level vision tasks into a single framework and proposes a novel image registration and fusion method, named Super Fusion. Specifically, we design a registration network to estimate bidirectional deformation fields to rectify geometric distortions of input images under the supervision of both photometric and end-point constraints. The registration and fusion are combined in a symmetric scheme, in which while mutual promotion can be achieved by optimizing the naive fusion loss, it is further enhanced by the mono-modal consistent constraint on symmetric fusion outputs. In addition, the image fusion network is equipped with the global spatial attention mechanism to achieve adaptive feature integration. Moreover, the semantic constraint based on the pre-trained segmentation model and Lovasz-Softmax loss is deployed to guide the fusion network to focus more on the semantic requirements of high-level vision tasks. Extensive experiments on image registration, image fusion,and semantic segmentation tasks demonstrate the superiority of our Super Fusion compared to the state-of-the-art alternatives.The source code and pre-trained model are publicly available at https://github.com/Linfeng-Tang/Super Fusion.

Total Variation Constrained Non-Negative Matrix Factorization for Medical Image Registration

This paper presents a novel medical image registration algorithm named total variation constrained graphregularization for non-negative matrix factorization(TV-GNMF).The method utilizes non-negative matrix factorization by total variation constraint and graph regularization.The main contributions of our work are the following.First,total variation is incorporated into NMF to control the diffusion speed.The purpose is to denoise in smooth regions and preserve features or details of the data in edge regions by using a diffusion coefficient based on gradient information.Second,we add graph regularization into NMF to reveal intrinsic geometry and structure information of features to enhance the discrimination power.Third,the multiplicative update rules and proof of convergence of the TV-GNMF algorithm are given.Experiments conducted on datasets show that the proposed TV-GNMF method outperforms other state-of-the-art algorithms.

Optical and SAR image registration based on improved nonsubsampled wavelet transform for sea islands

Homologous feature point extraction is a key problem in the optical and synthetic aperture radar （SAR） image registration for islands. A new feature point extraction method using a threshold shrink operator and non-subsampled wavelet transform （TSO-NSWT） for optical and SAR image registration was proposed. Moreover, the matching for this proposed feature was different from the traditional feature matching strategies and was performed using a similarity measure computed from neighborhood circles in low-frequency bands. Then, a number of reliably matched couples with even distributions were obtained, which assured the accuracy of the registration. Application of the proposed algorithm to SPOT-5 （multi-spectral） and YG-1 （SAR） images showed that a large number of accurately matched couples could be identified. Additionally, both of the root mean square error （RMSE） patterns of the registration parameters computed based on the TSO-NSWT algorithm and traditional NSWT algorithm were analyzed and compared, which further demonstrated the effectiveness of the proposed algorithm. The algorithm can supply the crucial step for island image registration and island recognition.

Two fully automated data-driven 3D whole-breast segmentation strategies in MRI for MR-based breast density using image registration and U-Net with a focus on reproducibility

Presence of higher breast density(BD)and persistence over time are risk factors for breast cancer.A quantitatively accurate and highly reproducible BD measure that relies on precise and reproducible whole-breast segmentation is desirable.In this study,we aimed to develop a highly reproducible and accurate whole-breast segmentation algorithm for the generation of reproducible BD measures.Three datasets of volunteers from two clinical trials were included.Breast MR images were acquired on 3T Siemens Biograph mMR,Prisma,and Skyra using 3D Cartesian six-echo GRE sequences with a fat-water separation technique.Two whole-breast segmentation strategies,utiliz-ing image registration and 3D U-Net,were developed.Manual segmentation was performed.A task-based analysis was performed:a previously developed MR-based BD measure,MagDensity,was calculated and assessed using automated and manual segmentation.The mean squared error(MSE)and intraclass correlation coefficient(ICC)between MagDensity were evaluated using the manual segmentation as a reference.The test-retest reproducibility of MagDensity derived from different breast segmentation methods was assessed using the difference between the test and retest measures(Δ_(2-1)),MSE,and ICC.The results showed that MagDensity derived by the registration and deep learning segmentation methods exhibited high concordance with manual segmentation,with ICCs of 0.986(95%CI:0.974-0.993)and 0.983(95%CI:0.961-0.992),respectively.For test-retest analysis,MagDensity derived using the regis-tration algorithm achieved the smallest MSE of 0.370 and highest ICC of 0.993(95%CI:0.982-0.997)when compared to other segmentation methods.In conclusion,the proposed registration and deep learning whole-breast segmentation methods are accurate and reliable for estimating BD.Both methods outperformed a previously developed algorithm and manual segmentation in the test-retest assessment,with the registration exhibiting superior performance for highly reproducible BD measurements.

Sonar Image Registration and Mosaic Based on Line Detection and Triangle Matching

Image registration is an important research topic in the field of computer vision,in which the registration and mosaic of side-scan sonar images is the keypoints of underwater navigation.However,the image registration method of keypoints is not suitable for sonar images which do not have obvious feature points.Therefore,a method of sonar-image registration and mosaic based on line segment extraction and triangle matching is proposed in this paper.Firstly,in order to extract features from sonar image,the LSD method is introduced to detect line feature from images,and line segments are filtered by the principle of attention;after that,triangles are formed from line segments,an image transformation matrix can be calculated through the heuristic greedy algorithm from these triangles;finally,images are merged based on the transformation information.On the basis of practical tests,it is found that,the feature extraction method used in this paper can better describe the outline of underwater terrain,and there is no obvious stitching gap between the result of sonar images stitched.Experimental results show that the proposed method is effective than the keypoints method of the registration and mosaic of sonar images.

Two new approaches for image registration based onspatial-temporal relationship

How to improve the probability of registration and precision of localization is a hard problem, which is desiderated to solve. The two basic approaches (normalized cross-correlation and phase correlation) for image registration are analysed, two improved approaches based on spatial-temporal relationship are presented. This method adds the correlation matrix according to the displacements in x- cirection and y- directions, and the registration pose is searched in the added matrix. The method overcomes the shortcoming that the probability of registration decreasing with area increasing owing to geometric distortion, improves the probability and the robustness of registration.

Automatic Image Registration Algorithm Based on Wavelet Transform

An automatic image registration approach based on wavelet transform is proposed. This proposed method utilizes multiscale wavelet transform to extract feature points. A coarse-to-fine feature matching method is utilized in the feature matching phase. A two-way matching method based on cross-correlation to get candidate point pairs and a fine matching based on support strength combine to form the matching algorithm. At last, based on an affine transformation model, the parameters are iteratively refined by using the leastsquares estimation approach. Experimental results have verified that the proposed algorithm can realize automatic registration of various kinds of images rapidly and effectively.

An image registration method based on multi-resolution morphology contour detection

Combined with the printing application,an image registration method based on the multi-resolution morphology contour detection was proposed.First,a direction based multi-resolution gray morphology in the scheme was proposed to realize the contour extraction.Then,based on the contour features,the subspace image registration was proposed to deal with issues of the computing complexity appeared in the traditional image registration methods.The proposed image registration was efficiently applied in the defect inspection of printing images.

Nonlinear Registration of Brain Magnetic Resonance Images with Cross Constraints of Intensity and Structure

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.

A comparison of cross-correlation-based and phasecorrelation-based image registration algorithms for optical coherence tomographic angiography

A large field of view is in high demand for disease diagnosis in clinical applications of optical coherence tomography(OCT)and OCT angiography(OCTA)imaging.Due to limits on the optical scanning range,the scanning speed,or the data processing speed,only a relatively small region could be acquired and processed for most of the current clinical OCT systems at one time and could generate a mosaic image of multiple adjacent small-region images with registration algorithms for disease analysis.In this work,we investigated performing cross-correlation(instead of phase-correlation)in the workflow of the Fourier–Mellin transform(FMT)method(called dual-cross-correlation-based translation and rotation registration,DCCTRR)for calculating translation and orientation offsets and compared its performance to the FMT method used on OCTA images alignment.Both phantom and in vivo experiments were implemented for comparisons,and the results quantitatively demonstrate that DCCTRR can align OCTA images with a lower overlap rate,which could improve the scanning efficiency of large-scale imaging in clinical applications.

GAN-DIRNet:A Novel Deformable Image Registration Approach for Multimodal Histological Images

Multi-modal histological image registration tasks pose significant challenges due to tissue staining operations causing partial loss and folding of tissue.Convolutional neural network(CNN)and generative adversarial network(GAN)are pivotal inmedical image registration.However,existing methods often struggle with severe interference and deformation,as seen in histological images of conditions like Cushing’s disease.We argue that the failure of current approaches lies in underutilizing the feature extraction capability of the discriminator inGAN.In this study,we propose a novel multi-modal registration approach GAN-DIRNet based on GAN for deformable histological image registration.To begin with,the discriminators of two GANs are embedded as a new dual parallel feature extraction module into the unsupervised registration networks,characterized by implicitly extracting feature descriptors of specific modalities.Additionally,modal feature description layers and registration layers collaborate in unsupervised optimization,facilitating faster convergence and more precise results.Lastly,experiments and evaluations were conducted on the registration of the Mixed National Institute of Standards and Technology database(MNIST),eight publicly available datasets of histological sections and the Clustering-Registration-Classification-Segmentation(CRCS)dataset on the Cushing’s disease.Experimental results demonstrate that our proposed GAN-DIRNet method surpasses existing approaches like DIRNet in terms of both registration accuracy and time efficiency,while also exhibiting robustness across different image types.

Research on Detection Technology of Micro-Components on Circuit Board Based on Digital Image Processing

Aiming at the stability of the circuit board image in the acquisition process,this paper realizes the accurate registration of the image to be registered and the standard image based on the SIFT feature operator and RANSAC algorithm.The device detection model and data set are established based on Faster RCNN.Finally,the number of training was continuously optimized,and when the loss function of Faster RCNN converged,the identification result of the device was obtained.

Fast Remote-Sensing Image Registration Using Priori Information and Robust Feature Extraction

In this paper, we propose a fast registration scheme for remote-sensing images for use as a fundamental technique in large-scale online remote-sensing data processing tasks. First, we introduce priori-information images,and use machine learning techniques to identify robust remote-sensing image features from state-of-the-art ScaleInvariant Feature Transform（SIFT） features. Next, we apply a hierarchical coarse-to-fine feature matching and image registration scheme on the basis of additional priori information, including a robust feature location map and platform imaging parameters. Numerical simulation results show that the proposed scheme increases position repetitiveness by 34%, and can speed up the overall image registration procedure by a factor of 7：47 while maintaining the accuracy of the image registration performance.

Phase Correlation Based Iris Image Registration Model

Iris recognition is one of the most reliable personal identification methods. In iris recognition systems, image registration is an important component. Accurately registering iris images leads to higher recognition rate for an iris recognition system. This paper proposes a phase correlation based method for iris image registration with sub-pixel accuracy. Compared with existing methods, it is insensitive to image intensity and can compensate to a certain extent the non-linear iris deformation caused by pupil movement. Experimental results show that the proposed algorithm has an encouraging performance.

Remote Sensing Image Registration Based on Improved KAZE and BRIEF Descriptor

Remote sensing image registration is still a challenging task owing to the significant influence of nonlinear differences between remote sensing images.To solve this problem,this paper proposes a novel approach with regard to feature-based remote sensing image registration.There are two key contributions:1)we bring forward an improved strategy of composite nonlinear diffusion filtering according to the scale factors in multi-scale space and 2)we design a gradually decreasing resolution of multi-scale pyramid space.And a binary code string is served as feature descriptors to improve matching efficiency.Extensive experiments of different categories of remote image datasets on feature extraction and feature registration are performed.The experimental results demonstrate the superiority of our proposed scheme compared with other classical algorithms in terms of correct matching ratio,accuracy and computation efficiency.

Two-step image registration by artificial immune system and chamfer matching

Image registration is the precondition and foundation in the fusion of multi-source image data. A two-step approach based on artificial immune system and chamfer matching to register images from different types of sensors is presented. In the first step, it extracts the large edges and takes chamfer distance between the input image and the reference image as similarity measure and uses artificial immune network algorithm to speed up the searching of the initial transformation parameters. In the second step, an area-based method is utilized to refine the initial transformation and enhance the registration accuracy. Experimental results show that the proposed approach is a promising method for registration of multi-sensor images.

Image registration based on matrix perturbation analysis using spectral graph

We present a novel perspective on characterizing the spectral correspondence between nodes of the weighted graph with application to image registration. It is based on matrix perturbation analysis on the spectral graph. The contribution may be divided into three parts. Firstly, the perturbation matrix is obtained by perturbing the matrix of graph model. Secondly, an orthogonal matrix is obtained based on an optimal parameter, which can better capture correspondence features. Thirdly, the optimal matching matrix is proposed by adjusting signs of orthogonal matrix for image registration. Experiments on both synthetic images and real-world images demonstrate the effectiveness and accuracy of the proposed method.