Efficacy of real-time artificial intelligence-aid endoscopic ultrasonography diagnostic system in discriminating gastrointestinal stromal tumors and leiomyomas: a multicenter diagnostic study.

Background: Gastrointestinal stromal tumors (GISTs) represent the most prevalent type of subepithelial lesions (SELs) with malignant potential. Current imaging tools struggle to differentiate GISTs from leiomyomas. This study aimed to create and assess a real-time artificial intelligence (AI) system using endoscopic ultrasonography (EUS) images to differentiate between GISTs and leiomyomas. Methods: The AI system underwent development and evaluation using EUS images from 5 endoscopic centers in China between January 2020 and August 2023. EUS images of 1101 participants with SELs were retrospectively collected for AI system development. A cohort of 241 participants with SELs was recruited for external AI system evaluation. Another cohort of 59 participants with SELs was prospectively enrolled to assess the real-time clinical application of the AI system. The AI system's performance was compared to that of endoscopists. This study is registered with Chictr.org.cn, Number ChiCT2000035787. Findings: The AI system displayed an area under the curve (AUC) of 0.948 (95% CI: 0.921-0.969) for discriminating GISTs and leiomyomas. The AI system's accuracy (ACC), sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) reached 91.7% (95% CI 87.5%-94.6%), 90.3% (95% CI 83.4%-94.5%), 93.0% (95% CI 87.2%-96.3%), 91.9% (95% CI 85.3%-95.7%), and 91.5% (95% CI 85.5%-95.2%), respectively. Moreover, the AI system exhibited excellent performance in diagnosing ≤20 mm SELs (ACC 93.5%, 95% CI 0.900-0.969). In a prospective real-time clinical application trial, the AI system achieved an AUC of 0.865 (95% CI 0.764-0.966) and 0.864 (95% CI 0.762-0.966) for GISTs and leiomyomas diagnosis, respectively, markedly surpassing endoscopists [AUC 0.698 (95% CI 0.562-0.834) for GISTs and AUC 0.695 (95% CI 0.546-0.825) for leiomyomas]. Interpretation: We successfully developed a real-time AI-assisted EUS diagnostic system. The incorporation of the real-time AI system during EUS examinations can assist endoscopists in rapidly and accurately differentiating various types of SELs in clinical practice, facilitating improved diagnostic and therapeutic decision-making. Funding: Science and Technology Commission Foundation of Shanghai Municipality, Science and Technology Commission Foundation of the Xuhui District, the Interdisciplinary Program of Shanghai Jiao Tong University and the Research Funds of Shanghai Sixth people's Hospital.

Locating X-Ray Coronary Angiogram Keyframes via Long Short-Term Spatiotemporal Attention With Image-to-Patch Contrastive Learning.

Locating the start, apex and end keyframes of moving contrast agents for keyframe counting in X-ray coronary angiography (XCA) is very important for the diagnosis and treatment of cardiovascular diseases. To locate these keyframes from the class-imbalanced and boundary-agnostic foreground vessel actions that overlap complex backgrounds, we propose long short-term spatiotemporal attention by integrating a convolutional long short-term memory (CLSTM) network into a multiscale Transformer to learn the segment- and sequence-level dependencies in the consecutive-frame-based deep features. Image-to-patch contrastive learning is further embedded between the CLSTM-based long-term spatiotemporal attention and Transformer-based short-term attention modules. The imagewise contrastive module reuses the long-term attention to contrast image-level foreground/background of XCA sequence, while patchwise contrastive projection selects the random patches of backgrounds as convolution kernels to project foreground/background frames into different latent spaces. A new XCA video dataset is collected to evaluate the proposed method. The experimental results show that the proposed method achieves a mAP (mean average precision) of 72.45% and a F-score of 0.8296, considerably outperforming the state-of-the-art methods. The source code is available at https://github.com/Binjie-Qin/STA-IPCon.

Robust PCA Unrolling Network for Super-Resolution Vessel Extraction in X-Ray Coronary Angiography.

Although robust PCA has been increasingly adopted to extract vessels from X-ray coronary angiography (XCA) images, challenging problems such as inefficient vessel-sparsity modelling, noisy and dynamic background artefacts, and high computational cost still remain unsolved. Therefore, we propose a novel robust PCA unrolling network with sparse feature selection for super-resolution XCA vessel imaging. Being embedded within a patch-wise spatiotemporal super-resolution framework that is built upon a pooling layer and a convolutional long short-term memory network, the proposed network can not only gradually prune complex vessel-like artefacts and noisy backgrounds in XCA during network training but also iteratively learn and select the high-level spatiotemporal semantic information of moving contrast agents flowing in the XCA-imaged vessels. The experimental results show that the proposed method significantly outperforms state-of-the-art methods, especially in the imaging of the vessel network and its distal vessels, by restoring the intensity and geometry profiles of heterogeneous vessels against complex and dynamic backgrounds. The source code is available at https://github.com/Binjie-Qin/RPCA-UNet.

An assessment of how the anterior cerebral artery anatomy impacts ACoA aneurysm formation based on CFD analysis.

OBJECTIVE: The aim of this study was to identify independent anatomic, morphologic and hemodynamic features of the ACoA (anterior communicating artery) complex that serve as risk factors for the occurrence of ACoA aneurysms. METHODS: Fifteen consecutive patients with 15 ACoA aneurysms were included. Computational fluid dynamics (CFD) simulations based on patient-specific models were carried out using 3D time-of-flight magnetic resonance angiography (3D-TOF-MRA) images. A reverse reconstruction technique was used to generate a pre-aneurysm vessel anatomy. Geometric parameters and hemodynamic changes were compared and evaluated. RESULTS: The overall prevalence of symmetric, dysplastic, and absent A1 segments were 53.3%, 26.7%, and 20%. The mean wall shear stress (WSS) of the absent group (AG) was significantly higher than that of the symmetric group (SG) and dysplastic group (DG). The absolute mean A1 artery flow rate (410.2 ± 88 versus 439.4 ± 101 mL/min; p = .45) of the aneurysm side was similar between the SG and DG but significantly higher in the AG (528.1 ± 77 mL/min; p < .05). The A1-A2 angles of the aneurysm side showed no significant differences among the 3 groups (p = .32). However, the mean A1-A2 angle on the aneurysm side was smaller than the contralateral A1-A2 angle (101.9 ± 9.1˚ versus 120.3 ± 7.7˚; p <.05). A regression analysis demonstrated that high WSS was significantly associated with a large A1-A2 ratio (R2=0.52; p <.05). CONCLUSIONS: ACoA aneurysms are a high-WSS pathology. Severe flow impingement and the anatomic vasculature structures play a role in triggering the occurrence of ACoA aneurysms.

Sequential vessel segmentation via deep channel attention network.

Accurately segmenting contrast-filled vessels from X-ray coronary angiography (XCA) image sequence is an essential step for the diagnosis and therapy of coronary artery disease. However, developing automatic vessel segmentation is particularly challenging due to the overlapping structures, low contrast and the presence of complex and dynamic background artifacts in XCA images. This paper develops a novel encoder-decoder deep network architecture which exploits the several contextual frames of 2D+t sequential images in a sliding window centered at current frame to segment 2D vessel masks from the current frame. The architecture is equipped with temporal-spatial feature extraction in encoder stage, feature fusion in skip connection layers and channel attention mechanism in decoder stage. In the encoder stage, a series of 3D convolutional layers are employed to hierarchically extract temporal-spatial features. Skip connection layers subsequently fuse the temporal-spatial feature maps and deliver them to the corresponding decoder stages. To efficiently discriminate vessel features from the complex and noisy backgrounds in the XCA images, the decoder stage effectively utilizes channel attention blocks to refine the intermediate feature maps from skip connection layers for subsequently decoding the refined features in 2D ways to produce the segmented vessel masks. Furthermore, Dice loss function is implemented to train the proposed deep network in order to tackle the class imbalance problem in the XCA data due to the wide distribution of complex background artifacts. Extensive experiments by comparing our method with other state-of-the-art algorithms demonstrate the proposed method's superior performance over other methods in terms of the quantitative metrics and visual validation. To facilitate the reproductive research in XCA community, we publicly release our dataset and source codes at https://github.com/Binjie-Qin/SVS-net.

Multi-Channel and Multi-Model-Based Autoencoding Prior for Grayscale Image Restoration.

Image restoration (IR) is a long-standing challenging problem in low-level image processing. It is of utmost importance to learn good image priors for pursuing visually pleasing results. In this paper, we develop a multi-channel and multi-model-based denoising autoencoder network as image prior for solving IR problem. Specifically, the network that trained on RGB-channel images is used to construct a prior at first, and then the learned prior is incorporated into single-channel grayscale IR tasks. To achieve the goal, we employ the auxiliary variable technique to integrate the higher-dimensional network-driven prior information into the iterative restoration procedure. In addition, according to the weighted aggregation idea, a multi-model strategy is put forward to enhance the network stability that favors to avoid getting trapped in local optima. Extensive experiments on image deblurring and deblocking tasks show that the proposed algorithm is efficient, robust, and yields state-of-the-art restoration quality on grayscale images.

[Deformable Registration Network Reinforced by Joint Saliency Map].

Image outliers such as missing correspondences and large local deformations break the one-to-one pixelwise mapping between target image and moving image to be registered. Both traditional registration methods and deep-learning based deformable image registration methods fail to tackle this problem. This paper proposed an unsupervised globalto-local deformable registration network reinforced by joint saliency map to accurately, robustly and fast address the problem. The global-to-local network divided the overall learning of a complex mapping of image registration into a simpler global mapping learning and local residual mapping. The joint saliency map of the two images to be registered bidirectionally reinforced the whole network's forward estimation and back-propagation with uncertainty modeling and context-aware intelligence. The experimental results confirm the proposed method's performance advantages over the state-of-the-arts registration methods in the challenges image registration with missing correspondences and large local deformations.

Phase asymmetry ultrasound despeckling with fractional anisotropic diffusion and total variation.

We propose an ultrasound speckle filtering method for not only preserving various edge features but also filtering tissue-dependent complex speckle noises in ultrasound images. The key idea is to detect these various edges using a phase congruence-based edge significance measure called phase asymmetry (PAS), which is invariant to the intensity amplitude of edges and takes 0 in non-edge smooth regions and 1 at the idea step edge, while also taking intermediate values at slowly varying ramp edges. By leveraging the PAS metric in designing weighting coefficients to maintain a balance between fractional-order anisotropic diffusion and total variation (TV) filters in TV cost function, we propose a new fractional TV framework to not only achieve the best despeckling performance with ramp edge preservation but also reduce the staircase effect produced by integral-order filters. Then, we exploit the PAS metric in designing a new fractional-order diffusion coefficient to properly preserve low-contrast edges in diffusion filtering. Finally, different from fixed fractional-order diffusion filters, an adaptive fractional order is introduced based on the PAS metric to enhance various weak edges in the spatially transitional areas between objects. The proposed fractional TV model is minimized using the gradient descent method to obtain the final denoised image. The experimental results and real application of ultrasound breast image segmentation show that the proposed method outperforms other state-of-the-art ultrasound despeckling filters for both speckle reduction and feature preservation in terms of visual evaluation and quantitative indices. The best scores on feature similarity indices have achieved 0.867, 0.844 and 0.834 under three different levels of noise, while the best breast ultrasound segmentation accuracy in terms of the mean and median dice similarity coefficient are 96.25% and 96.15%, respectively.

Accurate vessel extraction via tensor completion of background layer in X-ray coronary angiograms.

This paper proposes an effective method for accurately recovering vessel structures and intensity information from the X-ray coronary angiography (XCA) images of moving organs or tissues. Specifically, a global logarithm transformation of XCA images is implemented to fit the X-ray attenuation sum model of vessel/background layers into a low-rank, sparse decomposition model for vessel/background separation. The contrast-filled vessel structures are extracted by distinguishing the vessels from the low-rank backgrounds by using a robust principal component analysis and by constructing a vessel mask via Radon-like feature filtering plus spatially adaptive thresholding. Subsequently, the low-rankness and inter-frame spatio-temporal connectivity in the complex and noisy backgrounds are used to recover the vessel-masked background regions using tensor completion of all other background regions, while the twist tensor nuclear norm is minimized to complete the background layers. Finally, the method is able to accurately extract vessels' intensities from the noisy XCA data by subtracting the completed background layers from the overall XCA images. We evaluated the vessel visibility of resulting images on real X-ray angiography data and evaluated the accuracy of vessel intensity recovery on synthetic data. Experiment results show the superiority of the proposed method over the state-of-the-art methods.

Texture Variation Adaptive Image Denoising With Nonlocal PCA.

Image textures, as a kind of local variations, provide important information for the human visual system. Many image textures, especially the small-scale or stochastic textures, are rich in high-frequency variations, and are difficult to be preserved. Current state-of-the-art denoising algorithms typically adopt a nonlocal approach consisting of image patch grouping and group-wise denoising filtering. To achieve a better image denoising while preserving the variations in texture, we first adaptively group high correlated image patches with the same kinds of texture elements (texels) via an adaptive clustering method. This adaptive clustering method is applied in an over-clustering-and-iterative-merging approach, where its noise robustness is improved with a custom merging threshold relating to the noise level and cluster size. For texture-preserving denoising of each cluster, considering that the variations in texture are captured and wrapped in not only the between-dimension energy variations but also the within-dimension variations of PCA transform coefficients, we further propose a PCA-transform-domain variation adaptive filtering method to preserve the local variations in textures. Experiments on natural images show the superiority of the proposed transform-domain variation adaptive filtering to traditional PCA-based hard or soft threshold filtering. As a whole, the proposed denoising method achieves a favorable texture-preserving performance both quantitatively and visually, especially for irregular textures, which is further verified in camera raw image denoising.

Low-rank and sparse decomposition with spatially adaptive filtering for sequential segmentation of 2D+t vessels.

This letter proposes to extract contrast-filled vessels from overlapped noisy complex backgrounds in an x-ray coronary angiogram image sequence using low-rank and sparse decomposition. A refined vessel segmentation is finally achieved by implementing a radon-like feature filtering plus local-to-global adaptive thresholding to tackle the spatially varying noisy residuals in the extracted vessels. Based on real and synthetic XCA data, the experiment results demonstrate the superiority of the proposed method over the state-of-the-art methods.

Identifying epidermal growth factor receptor mutation status in patients with lung adenocarcinoma by three-dimensional convolutional neural networks.

OBJECTIVE:: Genetic phenotype plays a central role in making treatment decisions of lung adenocarcinoma, especially the tyrosine-kinase-inhibitors-sensitive mutations of the epidermal growth factor receptor (EGFR) gene. We constructed three-dimensional convolutional neural networks (CNN) to analyze underlying patterns in CT images that could indicate that EGFR gene mutation status but are invisible to human eyes. METHODS:: From 2012 to 2015, 503 Chinese patients with lung adenocarcinoma that had underwent surgery were included. Pathological types and EGFR mutation status were tested from surgical resections. EGFR mutations (exon 19 deletion or exon 21 L858R) were found in 215/345 (62.3%) and 91/158 (57.6%) patients in the training and independent validation set, respectively. CT images were taken before any invasive operation. The patients were randomly chosen to train the CNNs or validate the CNNs' performance. The performance was quantified using area under receiver operating characteristic curve (AUC), sensitivity, specificity, and accuracy. RESULTS:: The CNNs showed an AUC of 0.776 (range: 0.702-0.849, p< 0.0001) in the independent validation set and a fusion model of CNNs and clinical features (sex and smoking history) showed an AUC of 0.838 (range: 0.778-0.899, p< 0.0001), accuracy of 77.2%, sensitivity of 75.8% and specificity of 79.1% at the best diagnostic decision point. CONCLUSION:: The CNN exhibits potential ability to identify EGFR mutation status in patients with lung adenocarcinoma which might help make clinical decisions. ADVANCES IN KNOWLEDGE:: The CNN showed some diagnostic power and its performance could be further improved by increasing the training set, optimizing the network structure and training strategy. Medical image based CNN has the potential to reflect spatial heterogeneity.

A singular K-space model for fast reconstruction of magnetic resonance images from undersampled data.

Reconstructing magnetic resonance images from undersampled k-space data is a challenging problem. This paper introduces a novel method of image reconstruction from undersampled k-space data based on the concept of singularizing operators and a novel singular k-space model. Exploring the sparsity of an image in the k-space, the singular k-space model (SKM) is proposed in terms of the k-space functions of a singularizing operator. The singularizing operator is constructed by combining basic difference operators. An algorithm is developed to reliably estimate the model parameters from undersampled k-space data. The estimated parameters are then used to recover the missing k-space data through the model, subsequently achieving high-quality reconstruction of the image using inverse Fourier transform. Experiments on physical phantom and real brain MR images have shown that the proposed SKM method constantly outperforms the popular total variation (TV) and the classical zero-filling (ZF) methods regardless of the undersampling rates, the noise levels, and the image structures. For the same objective quality of the reconstructed images, the proposed method requires much less k-space data than the TV method. The SKM method is an effective method for fast MRI reconstruction from the undersampled k-space data. Graphical abstract Two Real Images and their sparsified images by singularizing operator.

Ultrasound Imaging Technologies for Breast Cancer Detection and Management: A Review.

Ultrasound imaging is a commonly used modality for breast cancer detection and diagnosis. In this review, we summarize ultrasound imaging technologies and their clinical applications for the management of breast cancer patients. The technologies include ultrasound elastography, contrast-enhanced ultrasound, 3-D ultrasound, automatic breast ultrasound and computer-aided detection of breast ultrasound. We summarize the study results seen in the literature and discuss their future directions. We also provide a review of ultrasound-guided, breast biopsy and the fusion of ultrasound with other imaging modalities, especially magnetic resonance imaging (MRI). For comparison, we also discuss the diagnostic performance of mammography, MRI, positron emission tomography and computed tomography for breast cancer diagnosis at the end of this review. New ultrasound imaging techniques, ultrasound-guided biopsy and the fusion of ultrasound with other modalities provide important tools for the management of breast patients.

Micro-Droplet Detection Method for Measuring the Concentration of Alkaline Phosphatase-Labeled Nanoparticles in Fluorescence Microscopy.

This paper developed and evaluated a quantitative image analysis method to measure the concentration of the nanoparticles on which alkaline phosphatase (AP) was immobilized. These AP-labeled nanoparticles are widely used as signal markers for tagging biomolecules at nanometer and sub-nanometer scales. The AP-labeled nanoparticle concentration measurement can then be directly used to quantitatively analyze the biomolecular concentration. Micro-droplets are mono-dispersed micro-reactors that can be used to encapsulate and detect AP-labeled nanoparticles. Micro-droplets include both empty micro-droplets and fluorescent micro-droplets, while fluorescent micro-droplets are generated from the fluorescence reaction between the APs adhering to a single nanoparticle and corresponding fluorogenic substrates within droplets. By detecting micro-droplets and calculating the proportion of fluorescent micro-droplets to the overall micro-droplets, we can calculate the AP-labeled nanoparticle concentration. The proposed micro-droplet detection method includes the following steps: (1) Gaussian filtering to remove the noise of overall fluorescent targets, (2) a contrast-limited, adaptive histogram equalization processing to enhance the contrast of weakly luminescent micro-droplets, (3) an red maximizing inter-class variance thresholding method (OTSU) to segment the enhanced image for getting the binary map of the overall micro-droplets, (4) a circular Hough transform (CHT) method to detect overall micro-droplets and (5) an intensity-mean-based thresholding segmentation method to extract the fluorescent micro-droplets. The experimental results of fluorescent micro-droplet images show that the average accuracy of our micro-droplet detection method is 0.9586; the average true positive rate is 0.9502; and the average false positive rate is 0.0073. The detection method can be successfully applied to measure AP-labeled nanoparticle concentration in fluorescence microscopy.

Fast single image super-resolution using estimated low-frequency k-space data in MRI.

PURPOSE: Single image super-resolution (SR) is highly desired in many fields but obtaining it is often technically limited in practice. The purpose of this study was to propose a simple, rapid and robust single image SR method in magnetic resonance (MR) imaging (MRI). METHODS: The idea is based on the mathematical formulation of the intrinsic link in k-space between a given (modulus) low-resolution (LR) image and the desired SR image. The method consists of two steps: 1) estimating the low-frequency k-space data of the desired SR image from a single LR image; 2) reconstructing the SR image using the estimated low-frequency and zero-filled high-frequency k-space data. The method was evaluated on digital phantom images, physical phantom MR images and real brain MR images, and compared with existing SR methods. RESULTS: The proposed SR method exhibited a good robustness by reaching a clearly higher PSNR (25.77dB) and SSIM (0.991) averaged over different noise levels in comparison with existing edge-guided nonlinear interpolation (EGNI) (PSNR=23.78dB, SSIM=0.983), zero-filling (ZF) (PSNR=24.09dB, SSIM=0.985) and total variation (TV) (PSNR=24.54dB, SSIM=0.987) methods while presenting the same order of computation time as the ZF method but being much faster than the EGNI or TV method. The average PSNR or SSIM over different slice images of the proposed method (PSNR=26.33 dB or SSIM=0.955) was also higher than the EGNI (PSNR=25.07dB or SSIM=0.952), ZF (PSNR=24.97dB or SSIM=0.950) and TV (PSNR=25.70dB or SSIM=0.953) methods, demonstrating its good robustness to variation in anatomical structure of the images. Meanwhile, the proposed method always produced less ringing artifacts than the ZF method, gave a clearer image than the EGNI method, and did not exhibit any blocking effect presented in the TV method. In addition, the proposed method yielded the highest spatial consistency in the inter-slice dimension among the four methods. CONCLUSIONS: This study proposed a fast, robust and efficient single image SR method with high spatial consistency in the inter-slice dimension for clinical MR images by estimating the low-frequency k-space data of the desired SR image from a single spatial modulus LR image.

[Poisson Noise Removal Using Patch-order Resampling PCA Algorithm].

The problem of Poisson denoising is common in various photon-limited imaging applications, especialy in low-light imaging, astronomy and nuclear medical applications. Due to the smal sample problem and the related insufficient self-similarity between patches of whole image, many denoising algorithms cannot obtain the favorable denoising performance. We propose patch-order resampling PCA algorithm for Poisson noise reduction. Firstly, we use the patch-ordered operations to sort the extracted image patches and exploit the bootstrap resampling method to resample the different blocks of spectral images to obtain more data matrix of image samples. Then, we select the patches with largest weights corresponding to the vectors of image samples data matrix as the most similar patches. Finaly, we use principal component analysis algorithm for processing the image to obtain the final denoised image. Experiments results show that the proposed method achieves excelent Poisson noise removal performance in the photon-limited images with smal sample problems.

Two-hierarchical nonnegative matrix factorization distinguishing the fluorescent targets from autofluorescence for fluorescence imaging.

BACKGROUND: Nonnegative matrix factorization (NMF) has been used in blind fluorescence unmixing for multispectral in-vivo fluorescence imaging, which decomposes a mixed source data into a set of constituent fluorescence spectra and corresponding concentrations. However, most classical NMF algorithms have ill convergence problems and they always fail to unmix multiple fluorescent targets from background autofluorescence for the sparse acquisition of multispectral fluorescence imaging, which introduces incomplete measurements and severe discontinuities in multispectral fluorescence emissions across the multiple spectral bands. METHODS: Observing the spatial distinction between the diffusive autofluorescence and the sparse fluorescent targets, we propose to separate the mixed sparse multispectral data into equality constrained two-hierarchical updating within NMF framework by dividing the concentration matrix of entire endmembers into two hierarchies: the fluorescence targets and the background autofluorescence. Specifically, when updating concentrations of multiple fluorescent targets in the two-hierarchical NMF, we assume that the concentration of autofluorescence is fixed and known, and vice versa. Furthermore, a sparsity constraint is imposed on the concentration matrix components of fluorescence targets only. RESULTS: Synthetic data sets, in vivo fluorescence imaging data are employed to demonstrate and validate the performance of our approach. The proposed algorithm can achieve more satisfying results of spectral unmixing and autofluorescence removal compared to other state-of-the-art methods, especially for the sparse multispectral fluorescence imaging. CONCLUSIONS: The proposed algorithm can successfully tackle the sparse acquisition and ill-posed problems in the NMF-based fluorescence unmixing through equality constraint along with partial sparsity constraint during two-hierarchical NMF optimization, at which fixing sparsity constrained target fluorescence can make the update of autofluorescence as accurate as possible and vice versa.

[Scale selection of local structures for medical image].

The scale of local structure is a key parameter in medical image registration. Unfortunately, no much attention has been paid to the scale selection for the local structures in the images. This paper proposes a data-driven scale selection method for local structures in the image. By using minimal description length criterion to maximize the posterior probability of local structure region with coherence constraint based on the Markov random field model, an optimal scale for each local structure, which is segmented with super-pixel representation, is assigned in terms of variance in a discrete anisotropic scale space. Therefore, the local structure's scale can be selected for further non-rigid medical image registration.