Efficient Registration Method Research Articles

Cultural Capital. Part two. Accounting

This part of the article dedicated to cultural value registration — a topical part of economic activity that is not currently solved. Author proposes an efficient registration method which, despite seemingly innumerable set of possible value parameters, allows to embrace more than 500 million of units within two years of operation, which includes consequent evaluation, employing hundreds of thousands of cultural field specialists in all areas of national culture, while increasing usefulness and demand of this profession.

Efficient low-overlapping point cloud registration based on two-stage network learning

ABSTRACT This letter introduces an efficient registration method for large-scale point clouds with low overlap, which adopts a two-stage transformation learning process. In the first stage, we integrate a point-voxel encoding structure with an attention mechanism to predict rotation and point-wise importance weight. In the second stage, we employ a PointNet-based structure to estimate translation using optimized data. By adopting this two-stage procedure, our model can effectively learn rotation and translation separately, thereby increasing flexibility. We conducted extensive experiments on two large-scale indoor datasets (3Dmatch and S3DIS) and an outdoor dataset (KITTI odometry). The results demonstrate that our method outperforms traditional and deep learning-based approaches, achieving state-of-the-art performance in registration. Furthermore, our method exhibits superior generalization capabilities, highlighting its effectiveness in various registration scenarios.

Registration‐enhanced multiple instance learning for cervical cancer whole slide image classification

AbstractHistopathological diagnosis is the golden standard for cancer diagnosis and grading. The past decade has witnessed great successes of vision‐based deep learning in computer‐aided cancer diagnosis. However, fully‐supervised methods require collecting large amounts of annotated data, which is a major challenge. In this research, we propose RMIL, a novel registration‐enhanced multiple instance learning pipeline, for cervical cancer biopsy grading of whole slide images as well as patch‐level regions. It requires only slide‐level annotations and a small amount of patch‐level annotations for training. To enhance it, an efficient registration method is implemented to generate more labeled patches with discriminative local features as data augmentations. In addition, attention mechanisms and self‐supervised learning are introduced to enhance feature extraction. Experiments showed that its classification accuracy is increased by 11.11%, compared to the vanilla multiple instance learning methods. The method has great potential for transferring to other scenarios.

Virtual trial assembly of large steel members with bolted connections based on point cloud data

Virtual trial assembly (VTA), an alternative to physical trial assembly, has been used in many projects owing to its advantages of low time consumption and low cost. However, the existing methods based on the terrestrial laser scanner (TLS) have low precision, are inefficient, and cannot guide rectification, making them impractical for large steel members with bolted connections. Therefore, this study introduces an automated approach to perform VTA using 3D laser scanning technology and building information modeling (BIM). In particular, a multi-scale data acquisition scheme that integrates a TLS and hand-held scanner is proposed to collect high accuracy point cloud data (PCD). Moreover, a rectification scheme based on particle swarm optimization is developed for gusset plates to guide rectification, and two efficient registration methods are established to align the scanned PCD with BIM model. The accuracy and feasibility of the proposed method are demonstrated through application to an actual suspension bridge.

Sparse-Based Domain Adaptation Network for OCTA Image Super-Resolution Reconstruction.

Retinal Optical Coherence Tomography Angiography (OCTA) with high-resolution is important for the quantification and analysis of retinal vasculature. However, the resolution of OCTA images is inversely proportional to the field of view at the same sampling frequency, which is not conducive to clinicians for analyzing larger vascular areas. In this paper, we propose a novel Sparse-based domain Adaptation Super-Resolution network (SASR) for the reconstruction of realistic [Formula: see text]/low-resolution (LR) OCTA images to high-resolution (HR) representations. To be more specific, we first perform a simple degradation of the [Formula: see text]/high-resolution (HR) image to obtain the synthetic LR image. An efficient registration method is then employed to register the synthetic LR with its corresponding [Formula: see text] image region within the [Formula: see text] image to obtain the cropped realistic LR image. We then propose a multi-level super-resolution model for the fully-supervised reconstruction of the synthetic data, guiding the reconstruction of the realistic LR images through a generative-adversarial strategy that allows the synthetic and realistic LR images to be unified in the feature domain. Finally, a novel sparse edge-aware loss is designed to dynamically optimize the vessel edge structure. Extensive experiments on two OCTA sets have shown that our method performs better than state-of-the-art super-resolution reconstruction methods. In addition, we have investigated the performance of the reconstruction results on retina structure segmentations, which further validate the effectiveness of our approach.

Simple and efficient registration of 3D point cloud and image data for an indoor mobile mapping system.

Registration of 3D lidar point clouds with optical images is critical in the combination of multisource data. Geometric misalignment originally exists in the pose data between lidar point clouds and optical images. To improve the accuracy of the initial pose and the applicability of the integration of 3D points and image data, we develop a simple but efficient registration method. We first extract point features from lidar point clouds and images: point features are extracted from single-frame lidar and point features are extracted from images using a classical Canny operator. The cost map is subsequently built based on Canny image edge detection. The optimization direction is guided by the cost map, where low cost represents the desired direction, and loss function is also considered to improve the robustness of the proposed method. Experiments show positive results.

Faster dense deformable image registration by utilizing both CPU and GPU.

.Purpose: Image registration is an important aspect of medical image analysis and a key component in many analysis concepts. Applications include fusion of multimodal images, multi-atlas segmentation, and whole-body analysis. Deformable image registration is often computationally expensive, and the need for efficient registration methods is highlighted by the emergence of large-scale image databases, e.g., the UK Biobank, providing imaging from 100,000 participants.Approach: We present a heterogeneous computing approach, utilizing both the CPU and the graphics processing unit (GPU), to accelerate a previously proposed image registration method. The parallelizable task of computing the matching criterion is offloaded to the GPU, where it can be computed efficiently, while the more complex optimization task is performed on the CPU. To lessen the impact of data synchronization between the CPU and GPU, we propose a pipeline model, effectively overlapping computational tasks with data synchronization. The performance is evaluated on a brain labeling task and compared with a CPU implementation of the same method and the popular advanced normalization tools (ANTs) software.Results: The proposed method presents a speed-up by factors of 4 and 8 against the CPU implementation and the ANTs software, respectively. A significant improvement in labeling quality was also observed, with measured mean Dice overlaps of 0.712 and 0.701 for our method and ANTs, respectively.Conclusions: We showed that the proposed method compares favorably to the ANTs software yielding both a significant speed-up and an improvement in labeling quality. The registration method together with the proposed parallelization strategy is implemented as an open-source software package, deform.

SAR Image Registration Based on ROEWA-Blocks and Multiscale Circle Descriptor

Given the imaging characteristics of synthetic aperture radar (SAR) images and the inherent speckle noise in them, scale-invariant feature transform based algorithms are unable to perform satisfactorily. To improve registration efficiency between SAR images, we propose a robust and efficient registration method with three main contributions. First, considering sudden dark patches appearing in SAR images, we propose the ratio of exponentially weighted average blocks to suppress the sudden dark patches and better adapt to different test images. This new operator called blocks of the ratio of exponentially weighted averages (ROEWA-B) divides the processing windows of ROEWA into blocks, which can not only reduce speckle noise but also retain more edge details compared to ROEWA when sudden dark patches appear. Second, for outlier removal, we present an approach using the minimum moment map to remove erroneous keypoints. Finally, based on the gradient location orientation histogram descriptor, we propose a novel multiscale circle descriptor, which combines scale change information to give weights to feature points at different scales. Experimental results for various thresholds and evaluations demonstrate the advantage and robustness of our method in registration.

Method for co-registration of high resolution 3Tmagnetic resonance imaging with standing 0.5T magnetic resonance imagingto investigate weight bearing joints

Method for co-registration of high resolution 3Tmagnetic resonance imaging with standing 0.5T magnetic resonance imagingto investigate weight bearing joints

Efficient diffeomorphic metric image registration via stationary velocity

Diffeomorphic image registration is a fundamental tool for MRI analysis, and fast growing image data demand highly efficient registration methods. Stationary velocity based method has much faster speed than non-stationary one in producing diffeomorphism, but how to achieve validity for large deformation while keeping its efficiency remains to be solved. To this end, we have proposed a new simplified model for optimal stationary velocity by representing temporal integration form of transformation variation via a single point model, e.g. the middle time point, according to mean value theorem and smoothness of transformation. This model can maintain the same registration accuracy as integral model but reduce the time cost a lot. It also shows as a better strategy than first order approximation model for large deformation mapping. Comparative study has also been conducted between this method and non-stationary approach on both synthesized and real brain images, and this approach demonstrated comparable whole brain mapping accuracy with much faster speed. Results showed efficacy of this approach in reducing complexity of stationary velocity based diffeomorphic registration while achieving validity for large deformation.

Practical optimal registration of terrestrial LiDAR scan pairs

Point cloud registration is a fundamental problem in 3D scanning. In this paper, we address the frequent special case of registering terrestrial LiDAR scans (or, more generally, levelled point clouds). Many current solutions still rely on the Iterative Closest Point (ICP) method or other heuristic procedures, which require good initializations to succeed and/or provide no guarantees of success. On the other hand, exact or optimal registration algorithms can compute the best possible solution without requiring initializations; however, they are currently too slow to be practical in realistic applications.Existing optimal approaches ignore the fact that in routine use the relative rotations between scans are constrained to the azimuth, via the built-in level compensation in LiDAR scanners. We propose a novel, optimal and computationally efficient registration method for this 4DOF scenario. Our approach operates on candidate 3D keypoint correspondences, and contains two main steps: (1) a deterministic selection scheme that significantly reduces the candidate correspondence set in a way that is guaranteed to preserve the optimal solution; and (2) a fast branch-and-bound (BnB) algorithm with a novel polynomial-time subroutine for 1D rotation search, that quickly finds the optimal alignment for the reduced set. We demonstrate the practicality of our method on realistic point clouds from multiple LiDAR surveys.

Automatic Registration of TLS-TLS and TLS-MLS Point Clouds Using a Genetic Algorithm.

Registration of point clouds is a fundamental issue in Light Detection and Ranging (LiDAR) remote sensing because point clouds scanned from multiple scan stations or by different platforms need to be transformed to a uniform coordinate reference frame. This paper proposes an efficient registration method based on genetic algorithm (GA) for automatic alignment of two terrestrial LiDAR scanning (TLS) point clouds (TLS-TLS point clouds) and alignment between TLS and mobile LiDAR scanning (MLS) point clouds (TLS-MLS point clouds). The scanning station position acquired by the TLS built-in GPS and the quasi-horizontal orientation of the LiDAR sensor in data acquisition are used as constraints to narrow the search space in GA. A new fitness function to evaluate the solutions for GA, named as Normalized Sum of Matching Scores, is proposed for accurate registration. Our method is divided into five steps: selection of matching points, initialization of population, transformation of matching points, calculation of fitness values, and genetic operation. The method is verified using a TLS-TLS data set and a TLS-MLS data set. The experimental results indicate that the RMSE of registration of TLS-TLS point clouds is 3~5 mm, and that of TLS-MLS point clouds is 2~4 cm. The registration integrating the existing well-known ICP with GA is further proposed to accelerate the optimization and its optimizing time decreases by about 50%.

Intelligent viewpoint selection for efficient CT to video registration in laparoscopic liver surgery

PurposeMinimally invasive surgery offers advantages over open surgery due to a shorter recovery time, less pain and trauma for the patient. However, inherent challenges such as lack of tactile feedback and difficulty in controlling bleeding lower the percentage of suitable cases. Augmented reality can show a better visualisation of sub-surface structures and tumour locations by fusing pre-operative CT data with real-time laparoscopic video. Such augmented reality visualisation requires a fast and robust video to CT registration that minimises interruption to the surgical procedure.MethodsWe propose to use view planning for efficient rigid registration. Given the trocar position, a set of camera positions are sampled and scored based on the corresponding liver surface properties. We implement a simulation framework to validate the proof of concept using a segmented CT model from a human patient. Furthermore, we apply the proposed method on clinical data acquired during a human liver resection.ResultsThe first experiment motivates the viewpoint scoring strategy and investigates reliable liver regions for accurate registrations in an intuitive visualisation. The second experiment shows wider basins of convergence for higher scoring viewpoints. The third experiment shows that a comparable registration performance can be achieved by at least two merged high scoring views and four low scoring views. Hence, the focus could change from the acquisition of a large liver surface to a small number of distinctive patches, thereby giving a more explicit protocol for surface reconstruction. We discuss the application of the proposed method on clinical data and show initial results.ConclusionThe proposed simulation framework shows promising results to motivate more research into a comprehensive view planning method for efficient registration in laparoscopic liver surgery.

Direct point-based registration for precise non-rigid surface matching using Student’s-t mixture model

Abstract One of the main challenges in the non-rigid surface matching is to match complex surfaces with absence of salient landmarks (marker-less) and salient structures (structure-less). We propose an accurate non-rigid surface registration method, called DSMM, to match complex surfaces based on a dense point-to-point correspondence alignment. The key idea of our approach is to model the correspondences on surfaces by using Student’s-t mixture model and represent local spatial structures via Dirichlet distribution and the directional springs. Firstly, we formulate the problem of alignment of two point sets as a probability density estimate, modeling one set as Student’s-t mixture model centroids, and the other one as observation data. We subsequently incorporate spatial representations of vertices on the surfaces into the prior probability of the finite Student’s-t mixture model by exploiting the Dirichlet distribution and Dirichlet law. We later explicitly add an additional structure regularization to get an approximate isometric and near-conformal transformation. Finally, we obtain closed-form solutions of registration parameters using Expectation Maximization (EM) framework, leading to a computationally efficient registration method. We compare DSMM with other state-of-the-art direct point-based non-rigid surface matching methods based on finite mixture models on artificial shapes with large deformation and real complex shapes from various segmented brain structures. DSMM demonstrates its statistical accuracy and robustness, outperforming the competing

Efficient registration of terrestrial LiDAR scans using a coarse-to-fine strategy for forestry applications

Terrestrial Laser Scanning (TLS) plays an increasing role in acquiring 3-dimensional forest structure information. Obstructions and stem density often need multiple scans, making the registration procedure vitally important. Since natural features are difficult to find, traditional registration methods often use artificial reflectors as tie points in forest areas. However, the placement of reflectors reduces the scanning efficiency, especially in forest stands with severe obstructions. In this paper, an efficient registration method using a coarse-to-fine strategy is proposed to overcome the challenge of placing artificial targets so as to improve the scanning efficiency. The coarse registration step utilizes the backsighting orientation approach to form a coarse alignment. Unlike other backsighting research used in urban areas, we solely rely on a TLS system with an internal compass and a built-in inclination sensor. In the fine registration step, stem center locations are calculated in each individual scan. Then, corresponding stem center pairs are found between adjacent scans with the help of the coarse registration results. They are used as tie points to perform a rigid-body transformation for the fine registration. The proposed coarse-to-fine registration method was tested at two forest sites in northeast China. The registration accuracy is approximately 1.5cm. Results also show that the diameters at breast height (DBHs) extracted from registered TLS data sets are highly correlated with field-measured DBHs (coefficient of determination (R2): 0.92, root mean square error (RMSE): 0.27cm). Considering both the scanning efficiency and accuracy, the proposed coarse-to-fine registration method provides a feasible and effective way for forest measurements.

An efficient strategy based on an individualized selection of registration methods. Application to the coregistration of MR and SPECT images in neuro-oncology

An efficient registration strategy is described that aims to help solve delicate medical imaging registration problems. It consists of running several registration methods for each dataset and selecting the best one for each specific dataset, according to an evaluation criterion. Finally, the quality of the registration results, obtained with the best method, is visually scored by an expert as excellent, correct or poor. The strategy was applied to coregister Technetium-99m Sestamibi SPECT and MRI data in the framework of a follow-up protocol in patients with high grade gliomas receiving antiangiogenic therapy. To adapt the strategy to this clinical context, a robust semi-automatic evaluation criterion based on the physiological uptake of the Sestamibi tracer was defined. A panel of eighteen multimodal registration algorithms issued from BrainVisa, SPM or AIR software environments was systematically applied to the clinical database composed of sixty-two datasets. According to the expert visual validation, this new strategy provides 85% excellent registrations, 12% correct ones and only 3% poor ones. These results compare favorably to the ones obtained by the globally most efficient registration method over the whole database, for which only 61% of excellent registration results have been reported. Thus the registration strategy in its current implementation proves to be suitable for clinical application.

NSCT edge Enhancement for SIFT key points extraction

Image registration is a key step for matching or mosaicing two or more images taken at different times, and/or with different sensors, hence the need for automatic methods arises. In this work, we present an efficient registration method based on the Non-Subsampled Contourlet Transform (NSCT) combined with the Scale Invariant Feature transform (SIFT) to extract robust local control points. Because NSCT is a shift-invariant multidirectional transform, it is used to extract edges at both spatial and directional resolutions. A comparative study is established between SIFT and NSCT-SIFT and experimental results show clearly that the proposed method (NSCT-SIFT) improves the registration accuracy.

An efficient registration method for partially overlapping images

In this paper, a method for accelerating registration of partially overlapping images is presented. Image registration is a fundamental problem in image processing and computer vision. Numerous studies have been done on this subject and can satisfy most of engineering application requirements. However, the registration processing is very time-consuming when the images are very large. So, time-saving methods are desirable.The proposed method, which is based on a feature-based method, concerns the computational cost of registration of partially overlapping images. The feature-based method generally contains four steps--features detecting, features matching, estimation of the homography and image warping and blending. The new method includes two stages. In the first stage, overlapping areas are rapidly estimated from the low-resolution correspondences of candidate images by a SIFT-based method. In the second stage, a homography between the image pairs is calculated by the same SIFT-based method, which involves only the overlapping areas between a target image and a reference image. Then the target image is warped and blended with the reference image. The experimental results demonstrated that our proposed method can reduce the computational cost to 10%∼30% of that of the SIFT-based homographic estimation (SHE) methods with little compromise in accuracy.

A computerized system to determine the provenance of finds in archaeological sites using acoustic signals

Some archaeological excavations require the accurate determination of the provenance of finds (three-dimensional location and orientation) for a subsequent spatial analysis. The traditional manual measurement using a grid reference is not a very efficient registration method. Some proposed computerized solutions based on total stations, photogrammetry and DGPS, are more effective but have some limitations. This paper presents a new acoustic localization system (3D-LOCUS) for measuring the three-dimensional position of finds in archaeological sites. The system can also be used for registering the size, shape and orientation of artifacts. 3D-LOCUS basically consists of a set of wireless rod-like pointing devices that are localized with a network of intelligent nodes installed above the excavation site. Archaeologists use the pointing device as a stylus to locate and outline the object under study. The main technological characteristics of the system are: omnidirectional wideband acoustic transducers, Bidirectional Time-Of-Arrival (BTOA) estimation, redundant ranging and robust trilateration. 3D-LOCUS achieves an accurate registration, even if used simultaneously to the digging labor, or in noisy or turbulent airflow conditions. We tested the system in the Gran Dolina archaeological site ( Atapuerca, Spain). The typical location accuracy is below 10 mm in natural conditions, and the achieved resolution is 5 mm, which allows us sketching objects with enough detail.

Registration with Uncertainties and Statistical Modeling of Shapes with Variable Metric Kernels

Registration and modeling of shapes are two important problems in computer vision and pattern recognition. Despite enormous progress made over the past decade, these problems are still open. In this paper, we advance the state of the art in both directions. First we consider an efficient registration method that aims to recover a one-to-one correspondence between shapes and introduce measures of uncertainties driven from the data which explain the local support of the recovered transformations. To this end, a free form deformation is used to describe the deformation model. The transformation is combined with an objective function defined in the space of implicit functions used to represent shapes. Once the registration parameters have been recovered, we introduce a novel technique for model building and statistical interpretation of the training examples based on a variable bandwidth kernel approach. The support on the kernels varies spatially and is determined according to the uncertainties of the registration process. Such a technique introduces the ability to account for potential registration errors in the model. Hand-written character recognition and knowledge-based object extraction in medical images are examples of applications that demonstrate the potentials of the proposed framework.