Rectangular Kernel Research Articles

Differentiation of high frequency near field versus far-field potentials in low voltage zones during sinus rhythm improves discrimination of local sites of termination during ventricular tachycardia

Abstract Background Ventricular tachycardia (VT) with structural heart disease (SHD) is often bad tolerated or presents with several inducible morphologies, this support mapping the arrhythmic substrate during sinus rhythm (SR). Poor discrimination between near-field (NF) and far-field (FF) electrogram (EGM) components in low voltage (LV) areas is the main limitation. The peak frequency (PF) associated with bipolar EGMs may distinguish NF from FF EGMs (Panel D). Objective To evaluate if PF combined with voltage mapping during SR enhances discrimination of the region of VT termination by radiofrequency application (RFa). Methods Bipolar activation and voltage maps were generated with a 16-pole grid catheter (HD Grid) during SR (Panel B). PF maps were retrospectively computed (Panel C). The maps were only analyzed if acute termination was achieved by focal RFa within a boundary diameter <1.5cm (Panel A). The mapped surface area (SA) was sub-regioned in termination (T-ZONE) and non-termination (NT-ZONE) zones. The T-ZONE surface was centered about the termination site and was allowed to extend to a rectangular kernel >0.5cm and <1.5cm across perpendicular axes, with the NT-ZONE encompassing the remainder mapped SA. Sensitivity (SE) and specificity (SP) of discrimination of the T-ZONE vs. NT-ZONE were measured according to the %SA covered at different LV thresholds (0.1-1.0mV in increments of 0.1mV - Panel F). LV high PF (LVHF) areas at different PF cutoffs from 150Hz (LVHF150) to 300Hz (LVHF300) were also assessed. Results 15 patients with SHD and sustained monomorphic VT were enrolled. 4 had 5 VT’s terminated by 1.4±0.9 focal RF lesions (60-74 y/o, 3 ischemic). All termination sites were located in a LVHF region mapped during SR (Panels B,C,E). The T-ZONE occupied 0.41±0.33cm /92.75±13.31cm global SA and consisted of 15±10 EGM points/1432±337 global points. T-ZONE voltage during SR (0.50±0.32mV) was significantly lower vs. global voltage (2.01±0.69mV, P<0.001), while T-ZONE PF was higher (324±51Hz) vs. global (270±37Hz, P<0.001). ROC discrimination of the T-ZONE yielded AUC’s of 0.79-LV alone, 0.87- LVHF150, 0.91-LVHF200 (Delta +12.2% vs LV alone, Panel F), 0.85-LVHF250 and 0.77-LVHF300. At a 0.7mV cutoff, LVHF200 yielded a balanced SE/SP of 0.89/0.87 vs. LV alone (0.90/0.63). Conclusion Assessment of regions with LVHF during SR improves SP for discrimination of sites of VT termination compared to LV alone.TV near field

Real-Time Assessment of Rodent Engagement Using ArUco Markers: A Scalable and Accessible Approach for Scoring Behavior in a Nose-Poking Go/No-Go Task.

In the field of behavioral neuroscience, the classification and scoring of animal behavior play pivotal roles in the quantification and interpretation of complex behaviors displayed by animals. Traditional methods have relied on video examination by investigators, which is labor-intensive and susceptible to bias. To address these challenges, research efforts have focused on computational methods and image-processing algorithms for automated behavioral classification. Two primary approaches have emerged: marker- and markerless-based tracking systems. In this study, we showcase the utility of "Augmented Reality University of Cordoba" (ArUco) markers as a marker-based tracking approach for assessing rat engagement during a nose-poking go/no-go behavioral task. In addition, we introduce a two-state engagement model based on ArUco marker tracking data that can be analyzed with a rectangular kernel convolution to identify critical transition points between states of engagement and distraction. In this study, we hypothesized that ArUco markers could be utilized to accurately estimate animal engagement in a nose-poking go/no-go behavioral task, enabling the computation of optimal task durations for behavioral testing. Here, we present the performance of our ArUco tracking program, demonstrating a classification accuracy of 98% that was validated against the manual curation of video data. Furthermore, our convolution analysis revealed that, on average, our animals became disengaged with the behavioral task at ∼75 min, providing a quantitative basis for limiting experimental session durations. Overall, our approach offers a scalable, efficient, and accessible solution for automated scoring of rodent engagement during behavioral data collection.

Data‐driven linear complexity low‐rank approximation of general kernel matrices: A geometric approach

SummaryA general, rectangular kernel matrix may be defined as where is a kernel function and where and are two sets of points. In this paper, we seek a low‐rank approximation to a kernel matrix where the sets of points and are large and are arbitrarily distributed, such as away from each other, “intermingled”, identical, and so forth. Such rectangular kernel matrices may arise, for example, in Gaussian process regression where corresponds to the training data and corresponds to the test data. In this case, the points are often high‐dimensional. Since the point sets are large, we must exploit the fact that the matrix arises from a kernel function, and avoid forming the matrix, and thus ruling out most algebraic techniques. In particular, we seek methods that can scale linearly or nearly linearly with respect to the size of data for a fixed approximation rank. The main idea in this paper is to geometrically select appropriate subsets of points to construct a low rank approximation. An analysis in this paper guides how this selection should be performed.

Systematic identification of low voltage-high frequency electrogram zones at sites of left atrial reentrant tachycardia termination

Abstract Funding Acknowledgements Type of funding sources: Private company. Main funding source(s): Abbott Laboratories Background Localization of the critical isthmus of conduction of LAMRT circuits within scar tissue is challenging (Panel A). Low voltage (LV) has been used as a marker for the arrhythmic substrate. However, voltage amplitudes due to near field (NF) vs. far-field (FF) electrogram (EGM) components are not clearly differentiated. The peak frequency (PF) associated with bipolar EGM’s is a novel parameter which may distinguish them (Panel B). Purpose To compare the ability of LV vs. LVHF to discriminate the LAMRT termination region Methods Bipolar voltage and activation maps were generated with a 16-pole grid catheter (Advisor HD Grid) during LAMRT (Panels A,C). Peak frequency maps were retrospectively computed (Panels B,D). LAMRT’s were included if acute termination could be achieved by focal RF lesion set restricted to a boundary <1.5cm diameter. The left atrial (LA) mapped surface was sub-regioned into the termination (T-ZONE) and non-termination (NT-ZONE) zones. The T-ZONE surface was centered about the site of termination and was allowed to extend to a rectangular kernel >0.5cm and <1.5cm across perpendicular axes, with the NT-ZONE encompassing the remainder of the mapped surface area (SA). Sensitivity (SE) and specificity (SP) of discrimination of the T-ZONE vs. NT-ZONE was measured according to the %T-ZONE vs. %NT-ZONE SA occupied at LV thresholds ranging from 0.1-1.0mV in increments of 0.1mV (Panel E). In addition to LV, LVHF at peak frequency cutoffs of 150Hz (LVHF150), 200Hz (LVHF200), 250Hz (LVHF250) and 300Hz (LVHF300) were also assessed for T-ZONE discrimination (Panels F,G). Results Of 16 consecutive patients with 24 ATs prospectively enrolled, 13 LAMRT’s from 13 patients met the termination criteria for inclusion. LAMRT termination consisted of a mean of 2.85±2.82 focal RF lesions. The T-ZONE occupied a mean SA of 0.82±0.30cm2 (41±19 EGM points) vs. 158.6±42.2cm2 (2967±1227 points) for the NT-ZONE. While the mean voltage from the T-ZONE was significantly lower vs. the NT-Zone (0.12±0.05mV vs. 0.60±0.28mV respectively), mean T-ZONE peak frequency of the was significantly higher vs. the NT-ZONE (345±76Hz vs. 255±34Hz). The maximum EGM duration in the T-ZONE was 172±63ms. ROC discrimination of the T-ZONE resulted in AUC’s of 0.69- LV, 0.76- LVHF150, 0.80- LVHF200, 0.81- LVHF250 and 0.76- LVHF300. LVHF200 and LVHF250 yielded the best improvements in AUC vs. LV unconstrained by high PF (11.3% and 12.2% respectively)(Panel G). LVHF200 with a cutoff of 0.2mV yielded a balanced SE/SP of 0.86/0.83 respectively, while LVHF250 with a cutoff of 0.3mV yielded a balanced SE/SP of 0.85/0.84. Conclusion During AT, systematic identification of EGM regions of high frequency within low voltage can significantly improve discrimination of the site of AT. termination, beyond.

Peak frequency mapping to differentiate near-field from far-field electrograms for ventricular tachycardia ablation: initial results

Abstract Funding Acknowledgements Type of funding sources: Private company. Main funding source(s): Abbott Laboratories Background Ablation of ventricular tachycardia (VT) in patients with structural heart disease is challenging. Localization of the critical isthmus of conduction may be difficult because is commonly located within low voltage (LV) areas where far-field (FF) and near-field (NF) electrograms (EGMs) may be present and merged. The peak frequency (PF) associated with bipolar EGMs is a novel parameter which may serve to distinguish NF from FF EGM components. The potential value of PF to identify the critical isthmus of conduction in this setting has never been evaluated. Purpose To compare the ability of LV vs. LVHF to discriminate the VT termination region Methods Bipolar voltage and activation maps were generated with a 16-pole grid catheter (Advisor HD Grid) during VT. Peak frequency maps were retrospectively computed. VTs were included if acute termination could be achieved through application of a focal RF lesions set restricted to a boundary <1.5cm diameter. The ventricular LV mapped surface was sub-regioned into the termination (T-ZONE) and non-termination (NT-ZONE) zones. The T-ZONE surface was centered about the site of termination and was allowed to extend to a rectangular kernel >0.5cm and <1.5cm across perpendicular axes, with the NT-ZONE encompassing the remainder of the mapped surface area (SA). Sensitivity (SE) and specificity (SP) of discrimination of the T-ZONE vs. NT-ZONE was measured according to the %T-ZONE vs. %NT-ZONE SA occupied at LV thresholds ranging from 0.1-1.0mV in increments of 0.1mV (Panel E). In addition to LV, LVHF at peak frequency cutoffs of 200Hz (LVHF200), 250Hz (LVHF250), 300Hz (LVHF300) and 400Hz (LVHF400) were also assessed for T-ZONE discrimination. Results 8 patients with sustained monomorphic VT and structural heart disease (ischemic and non-ischemic) were enrolled. All of them had VT terminated with limited ablation by focal ablation at a discrete region. All areas of VT termination were located in a LV and high PF zone (figure 1). The best area under the curve to identify them was using the 250 Hz cuttoff (AUC 0.91) which was also better that using omnipolar optimized voltage alone (AUC 0.84) (Figure 2). Conclusion PF mapping apparently identifies slow conduction isthmus and provides complementary information to bipolar voltage mapping which can be displayed with emphasis on the electranatomical map.

AB-452674-1 HIGH-FREQUENCY ELECTROGRAM CHARACTERISTICS IN REGIONS OF LOW VOLTAGE ACCURATELY DISCRIMINATE SITES OF LEFT ATRIAL TACHYCARDIA TERMINATION

Localization of the critical isthmus of conduction of left atrial (LA) macroreentrant tachycardia (LAMRT) circuits within scar tissue is challenging (Panel A). Low voltage (LV) has been used as a marker for the arrhythmic substrate. However, voltage amplitudes due to near-field (NF) vs. far-field (FF) electrogram (EGM) components are not clearly differentiated. The peak frequency (PF) associated with bipolar EGM’s is a novel parameter which may distinguish them (Panel B). To compare the ability of LV vs. LV with high PF (LVHF) to discriminate the LAMRT termination region. Bipolar voltage and activation maps were generated with a 16-pole grid catheter (HD-Grid) during LAMRT (Panels A,B). PF maps were retrospectively computed (Panel C). LAMRT’s were only included if terminated by focal RF lesions (boundary <1.5cm diameter). The LA mapped surface was sub-regioned into the termination (T-ZONE) and non-termination (NT-ZONE) zones. The T-ZONE was centered about the termination site and allowed to extend to a rectangular kernel >0.5cm and <1.5cm across perpendicular axes. The NT-ZONE encompassed the remainder of the mapped surface area (SA). Sensitivity (SE) and specificity (SP) of T-ZONE vs. NT-ZONE discrimination was measured according to the %T-ZONE vs. %NT-ZONE SA occupied at LV thresholds (range 0.1-1.0mV) in 0.1mV increments (Panel E). LVHF at PF cutoffs of 150Hz (LVHF150), 200Hz (LVHF200), 250Hz (LVHF250) and 300Hz (LVHF300) were also assessed for TZONE discrimination (Panel F). 16 consecutive patients with 24 ATs were prospectively enrolled and 13 LAMRT’s were terminated by 2.85±2.82 focal RF lesions. The T-ZONE occupied 0.82±0.30cm2 (41±19 EGM points) vs. 158.6±42.2cm2 (2967±1227 points) for the NT-ZONE. T-ZONE voltage was significantly lower vs. the NT-ZONE (0.12±0.05mV vs. 0.60±0.28mV, P<0.001 respectively). T-ZONE PF was significantly higher than the NT-ZONE PF (345±76Hz vs. 255±34Hz, P<0.001). ROC discrimination of the T-ZONE resulted in AUC’s of 0.69- LV, 0.86- LVHF150, 0.88- LVHF200, 0.86- LVHF250 and 0.78- LVHF300. LVHF200 yielded the best improvement of AUC vs. LV unconstrained by high PF (19.1%)(Panel E). At a 0.2 mV cutoff, LVHF200 yielded a balanced SE/SP of 0.86/0.83 respectively vs. SE/SP of 0.90/0.54 for LV alone. During AT, systematic identification of EGM regions of high frequency within LV can significantly improve the specificity of discrimination of the site of AT termination.

PO-04-228 DIFFERENTIATION OF HIGH FREQUENCY NEAR FIELD VS. FAR-FIELD POTENTIALS IN LOW VOLTAGE ZONES DURING SINUS RHYTHM IMPROVES DISCRIMINATION OF LOCAL SITES OF TERMINATION DURING VT

Ventricular tachycardia (VT) in patients with structural heart disease (SHD) is often poorly tolerated or presents with several inducible morphologies. These factors support mapping the arrhythmic substrate during sinus rhythm (SR). However, poor discrimination between near-field (NF) and far-field (FF) electrogram (EGM) components in low voltage (LV) areas is the main limitation. The peak frequency (PF) associated with bipolar EGMs may distinguish NF from FF EGMs (Panel D). To evaluate if PF combined with voltage mapping during SR enhances discrimination of the region of VT termination by radiofrequency application (RFa). Bipolar activation and voltage maps were generated with a 16-pole grid catheter (HD Grid) during SR (Panel B). PF maps were retrospectively computed (Panel C). The maps were only analysed if acute termination was achieved by focal RFa within a boundary diameter<1.5cm (panel A). The mapped surface area (SA) was sub-regioned into the termination (T-ZONE) and non-termination (NT-ZONE) zones. The T-ZONE surface was centered about the termination site and was allowed to extend to a rectangular kernel >0.5cm and <1.5cm across perpendicular axes, with the NT-ZONE encompassing the remainder mapped SA. Sensitivity (SE) and specificity (SP) of discrimination of the T-ZONE vs. NT-ZONE were measured according to the %SA covered at different LV thresholds (0.1-1.0mV in increments of 0.1mV - Panel F). LV high PF (LVHF) areas at different PF cutoffs from 150Hz (LVHF150) to 300Hz (LVHF300) were also assessed. 15 patients with SHD and sustained monomorphic VT were enrolled. Four (60-74 years, all male, 3 ischemic) of them had 5 VT’s terminated by 1.4±0.9 focal RF lesions. All termination sites were located in a LVHF region mapped during SR (Panels B,C,E). The T-ZONE occupied 0.41±0.33cm2/92.75±13.31cm2 global SA, and consisted of 15±10 EGM points/1432±337 global points. T-ZONE voltage during SR (0.50±0.32mV) was significantly lower vs. global voltage (2.01±0.69mV, P<0.001), while T-ZONE PF was higher (324±51Hz) vs. global (270±37Hz, P<0.001). ROC discrimination of the T-ZONE yielded AUC’s of 0.79-LV alone, 0.87-LVHF150, 0.91-LVHF200 (Delta +12.2% vs LV alone, Panel F), 0.85-LVHF250 and 0.77-LVHF300. At a 0.7mV cutoff, LVHF200 yielded a balanced SE/SP of 0.89/0.87 vs. LV alone (0.90/0.63). Assessment of regions with LV and high PF during SR improves specificity for discrimination of sites of VT termination compared to LV alone.

S-ACF: a selective estimator for the autocorrelation function of irregularly sampled time series

ABSTRACT We present a generalized estimator for the autocorrelation function, S-ACF, which is an extended version of the standard estimator of the autocorrelation function (ACF). S-ACF is a versatile definition that can robustly and efficiently extract periodicity and signal shape information from a time series, independent of the time sampling and with minimal assumptions about the underlying process. Calculating the autocorrelation of irregularly sampled time series becomes possible by generalizing the lag of the standard estimator of the ACF to a real parameter and introducing the notion of selection and weight functions. We show that the S-ACF reduces to the standard ACF estimator for regularly sampled time series. Using a large number of synthetic time series, we demonstrate that the performance of the S-ACF is as good or better than commonly used Gaussian and rectangular kernel estimators, and is comparable to a combination of interpolation and the standard estimator. We apply the S-ACF to astrophysical data by extracting rotation periods for the spotted star KIC 5110407, and compare our results to Gaussian process (GP) regression and Lomb–Scargle (LS) periodograms. We find that the S-ACF periods typically agree better with those from GP regression than from LS periodograms, especially in cases where there is evolution in the signal shape. The S-ACF has a wide range of potential applications and should be useful in quantitative science disciplines where irregularly sampled time series occur. A python implementation of the S-ACF is available under the MIT license.

CNN parameter design based on fault signal analysis and its application in bearing fault diagnosis

As a representative deep learning network, Convolutional Neural Network (CNN) has been extensively used in bearing fault diagnosis and many good results have been reported. In Prognostics and Health Management (PHM) field, the CNN’s input size is usually designed as a 1D vector or 2D square matrix, and the convolution kernel size is also defined as a square shape like 3 × 3 and 5 × 5, which are directly adopted from the image recognition. Though satisfying results can be obtained, CNN with such parameter specifications is not optimal and efficient. To this end, this paper elaborated the physical characteristics of bearing acceleration signals to guide the CNN design. First, the fault period under different fault types and shaft rotation frequency were used to determine the size of CNN’s input. Next, an exponential function was involved in fitting the envelope of decaying acceleration signal during each fault period, and signal length within different decaying ratios was used to define the CNN’s kernel size. Finally, the designed CNN was validated with the Case Western Reserve University bearing dataset and Paderborn University bearing dataset. Results confirm that the physics-guided CNN (PGCNN) with rectangular input shape and rectangular convolution kernel works better than the baseline CNN with higher accuracy and smaller uncertainty. The feasibility of designing CNN parameters with physics-guided rules derived from bearing fault signal analysis has also been verified.

Instance-Aware Spatial-Frequency Feature Fusion Detector for Oriented Object Detection in Remote-Sensing Images

In recent years, fusing multi-type features poses great potential for oriented object detection (OOD) in remote sensing images (RSIs). Due to the inexplicit operation of modeling orientation variations, convolutional neural networks (CNNs) are difficult to perceive objects under different transformations (angles and scales). In this paper, we propose a novel instance-aware spatial-frequency feature fusion detector (SFFD) for oriented object detection in remote sensing images. First, a layer-wise frequency-domain analysis (L-FDA) module is built along with CNN layers to extract frequency features. Getting rid of the constrains such as horizontal rectangular kernel in CNNs, our L-FDA possesses outstanding ability of locating mutational signals from frequency space. These mutational signals record the scale and angle information of the oriented instances in images. Subsequently, CNN and frequency features are sent into RoI Pooling layer to obtain multi-type instance-level RoI features. Moreover, the proposed instance-aware cross feature fusion (CFF) module explores the interaction between these diverse features which provides an explicit indicator to compensate the orientation information ignored by instance-level CNN features. Finally, our SFFD unifies the proposed L-FDA module and CFF module into the detection network to localize oriented instances in RSIs. We compare our method with many state-of-the-art methods on DOTA, HRSC2016, and NWPU VHR-10 datasets. Experimental results verify the validity of modeling instance-level object relations from frequency-domain and CNNs for OOD.

Inverse free reduced universum twin support vector machine for imbalanced data classification

Imbalanced datasets are prominent in real-world problems. In such problems, the data samples in one class are significantly higher than in the other classes, even though the other classes might be more important. The standard classification algorithms may classify all the data into the majority class, and this is a significant drawback of most standard learning algorithms, so imbalanced datasets need to be handled carefully. One of the traditional algorithms, twin support vector machines (TSVM), performed well on balanced data classification but poorly on imbalanced datasets classification. In order to improve the TSVM algorithm’s classification ability for imbalanced datasets, recently, driven by the universum twin support vector machine (UTSVM), a reduced universum twin support vector machine for class imbalance learning (RUTSVM) was proposed. The dual problem and finding classifiers involve matrix inverse computation, which is one of RUTSVM’s key drawbacks. In this paper, we improve the RUTSVM and propose an improved reduced universum twin support vector machine for class imbalance learning (IRUTSVM). We offer alternative Lagrangian functions to tackle the primal problems of RUTSVM in the suggested IRUTSVM approach by inserting one of the terms in the objective function into the constraints. As a result, we obtain new dual formulation for each optimization problem so that we need not compute inverse matrices neither in the training process nor in finding the classifiers. Moreover, the smaller size of the rectangular kernel matrices is used to reduce the computational time. Extensive testing is carried out on a variety of synthetic and real-world imbalanced datasets, and the findings show that the IRUTSVM algorithm outperforms the TSVM, UTSVM, and RUTSVM algorithms in terms of generalization performance.

Adaptive Modulation and Rectangular Convolutional Network for Stereo Image Super-Resolution

• We propose a novel network for better stereo image super-resolution. • We propose an adaptive modulation alignment mechanism to modulate the aligned features. • We add rectangular convolutional layers to expand the receptive field along the disparity epipolar line . • Experimental results show that our method outperforms other state-of-the-art methods on the three stereo benchmarks. Deep learning based stereo image super-resolution algorithm can make use of the additional information in the stereo image pairs to improve the quality of the reconstructed images. However, extracting the similarity information of one image to another precisely is challenging when partially occluded regions exist in stereo images. In this paper, we propose an adaptive modulation alignment mechanism to modulate the aligned features calculated from the parallax attention mechanism and effectively deal with the inaccuracy caused by occlusion. Furthermore, because disparities of stereo image exist only along the epipolar line, we use rectangular convolution kernel in some convolution layers , to extend the receptive field horizontally. Finally, experimental results demonstrate that our method achieves state-of-the-art performance on the Middlebury, KITTI 2012 and KITTI 2015 stereo benchmarks.

TSVM-M[formula omitted]: Twin support vector machine based on multi-order moment matching for large-scale multi-class classification

For multi-class classification, many existing methods, such as multiple weighted linear loss twin support vector machine (MWLTSVM), construct multiple decision hyperplanes by minimizing the positive points loss’s first-order moment (mean), which may lead to sensitivity to outliers. Also, when faced with a large-scale classification problem, how to speed up the process of solving the optimization model is also a challenge. An alternative is to use rectangular kernel technology (RKT) to reduce computational complexity. However, RKT is based on the uniform point selection method, which can be proven to be ineffective in improving classifier performance. To address these problems, a novel classifier under the structure of “one-versus-rest” for multi-class classification is proposed in this paper, named twin support vector machine based on multi-order moment matching (TSVM-M3). When constructing the decision hyperplanes, TSVM-M3 takes the first-order and second-order moments (mean and variance) of positive points loss into consideration and implements this by introducing an adjusting factor into the objective function. A theoretical analysis of the robustness of the proposed TSVM-M3 is also provided. Meanwhile, a novel RKT based on the density-dependent data selection method is proposed for large-scale classification. We demonstrate that the proposed RKT can benefit from reducing modeling error. Experimental results show the effectiveness of the proposed TSVM-M3.

On the shape of convolution kernels in MRI reconstruction: Rectangles versus ellipsoids.

Many MRI reconstruction methods (including GRAPPA, SPIRiT, ESPIRiT, LORAKS, and convolutional neural network [CNN] methods) involve shift-invariant convolution models. Rectangular convolution kernel shapes are often chosen by default, although ellipsoidal kernel shapes have potentially appealing theoretical characteristics. In this work, we systematically investigate the differences between different kernel shape choices in several contexts. It is well-understood that a rectangular region of k-space is associated with anisotropic spatial resolution, while ellipsoidal regions can be associated with more isotropic resolution. Further, for a fixed spatial resolution, ellipsoidal kernels are associated with substantially fewer parameters than rectangular kernels. These characteristics suggest that ellipsoidal kernels may have certain advantages over rectangular kernels. We used real retrospectively undersampled k-space data to empirically study the characteristics of rectangular and ellipsoidal kernels in the context of seven methods (GRAPPA, SPIRiT, ESPIRiT, SAKE, LORAKS, AC-LORAKS, and CNN-based reconstructions). Empirical results suggest that both kernel shapes can produce reconstructed images with similar error metrics, although the ellipsoidal shape can often achieve this with reduced computation time and memory usage and/or fewer model parameters. Ellipsoidal kernel shapes may offer advantages over rectangular kernel shapes in various MRI applications.

Spine Image Segmentation Based on U-Net and Atrous spatial pyramid pooling

In the past ten years, deep learning has achieved remarkable results in the area of natural image segmentation, and has gradually turned to the field of medical image segmentation. The precise segmentation of spine images can be used for early screening of spondylopathy, which is convenient for early detection and treatment of patients. Aiming at the segmentation of the spine image by the U-Net network, which structure will lead to large model calculation, network overfitting, image size, noise information and other issues. This paper introduces a new method based on spatial pyramid pooling module ASpp on U-Net network and applies the proposed network structure to the segmentation of the spine image. The model is first enhanced by the Spp module and Densenet data; secondly, the U-Net encoding and decoding structure is adopted, and the deep separable convolution is used. This method greatly reduces the complexity and computation of the model, and uses a rectangular convolution kernel to increase the computation of the model in a small amount. Based on the volume, the receptive field of the convolution operation is expanded; finally, in order to effectively enhance the segmentation area of the image, the SE-Net attention mechanism module is added to the lateral connection. The method proposed in this paper conducts ablation experiments on a public dataset of spine images and compares it with the U-Net network. The method proposed IOU in this paper can reach 72%, and the F1-score can reach 0.92. The comparison of the experimental results of spine medical image datasets shows that the method proposed in this paper can effectively improve the accuracy and accuracy of spine image segmentation.

Conditional adversarial consistent identity autoencoder for cross-age face synthesis

Learning-based face aging/rejuvenation has witnessed rapid progress in recent years. However, existing methods still suffer from the loss of personalized identity information when synthesizing cross-age faces. In this paper, we propose a Conditional Adversarial Consistent Identity AutoEncoder (CACIAE) to revisit this problem. Firstly, a Res-Encoder is designed to better generate powerful face representation. Secondly, the rectangular kernel is introduced into the encoder to make full use of horizontal continuous characteristic information of faces and to make the synthetic face images more natural. Thirdly, a novel consistent identity loss is proposed to learn more face details and produce more natural identity-preserving images. Further, two discriminators are designed to enforce the generator to generate more realistic and more age-accurate images. Experimental results prove the effectiveness of the proposed method, both qualitatively and quantitatively. The code is available at https://github.com/XH-B/CACIAE.

Farklı Dağılımlar ve Farklı Çekirdek Fonksiyonları için ROC Eğrisi Tahmin Yöntemlerinin Performanslarının İncelenmesi

Objective: Receiver operating characteristic (ROC) curve is a statistical method used to examine the actual effectiveness of a diagnostic test or a biomarker in a comprehensive and reliable way. Several methods have been proposed to estimate ROC curve properly. The aim of the present study is to compare recent ROC curve estimation methods for different distribution and sample sizes. Material and Methods: Log-concave density and smooth log-concave density estimate based ROC curve estimation, kernel based ROC curve estimation with Gaussian, Epanechnikov, rectangular, triangular kernels, and binormal ROC estimation methods were compared for different simulation scenarios. Results: The ROC curve estimation methods based on kernel estimates gave their best performances when the biomarker values of non-diseased group are normal but the biomarker values of the diseased group are right-skewed, with a notable difference from other methods. Epanechnikov and rectangular kernel methods yielded better performance than other kernel methods in small sample sizes; but this difference disappeared as the sample size increased. The methods based on kernel or log-concave density estimate gave their worst results for the simulation scenario where the data were nonnormal but symmetric. Conclusion: The performances of the other methods examined in the study exceeded the performance of the binormal method in highly skewed data in both groups and when the distribution of diseased and non-diseased populations were right-skewed and normal, respectively.

Automatic detection of pulmonary nodules in CT images based on 3D Res-I network

It is difficult for the existing detection methods of the pulmonary nodules to take into account the global and local features simultaneously. It will lead to over-fitting and lower sensitivity since the extracted features of 3D pulmonary nodules is too complex. To solve these problems, a model based an improved 3D residual structure (3D Res-I) was proposed to detect pulmonary nodules. In the model, the basic residual structure is improved by using rectangular convolution kernel, grouping convolution and pre-activation. Rectangular convolution kernel expands the receptive filed of the convolution, which effectively takes into account the global and local features of the pulmonary nodules. Grouping convolution reduces the computational cost of the model. Pre-activation operation alleviates over-fitting phenomenon. 3D Res-I structure is combined with the improved U-Net network as the feature extraction network of Faster R-CNN. The experimental results on LUNA16 dataset show that the proposed model improves the detection accuracy of pulmonary nodules and reduces the average number of false positives and the size of the generated model.

A reduced universum twin support vector machine for class imbalance learning

In most of the real world datasets, there is an imbalance in the number of samples belonging to different classes. Various pattern classification problems such as fault or disease detection involve class imbalanced data. The support vector machine (SVM) classifier becomes biased towards the majority class due to class imbalance. Moreover, in the existing SVM based techniques for class imbalance, there is no information about the distribution of data. Motivated by the idea of prior information about data distribution, a reduced universum twin support vector machine for class imbalance learning (RUTSVM-CIL) is proposed in this paper. For the first time, universum learning is incorporated with SVM to solve the problem of class imbalance. Oversampling and undersampling of data is performed to remove the imbalance in the classes. The universum data points are used to give prior information about the data. To reduce the computation time of our universum based algorithm, we use a small sized rectangular kernel matrix. The reduced kernel matrix needs less storage space, and thus applicable for large scale imbalanced datasets. Comprehensive experimentation is performed on various synthetic, real world and large scale imbalanced datasets. In comparison to the existing approaches for class imbalance, the proposed RUTSVM-CIL gives better generalization performance for most of the benchmark datasets. Also, the computation cost of RUTSVM-CIL is very less, making it suitable for real world applications.

Annual and seasonal cycles of CO2 and CH4 in a Mediterranean Spanish environment using different kernel functions

This paper is based on CO2 and CH4 semi-hourly mole fraction measurements obtained at the Low Atmosphere Research Centre (CIB) between 2010 and 2016 using a Picarro G1301 analyser. The main aims of the study were to examine the temporal variation of CO2 and CH4 by using six different kernel functions, and to study the suitability of these functions to the dataset. The method used for the current study was based on experimental contour plots of R2 values in order to simultaneously determine the bandwidths of kernel functions for the long-term and short-term. An Epanechnikov, a Gaussian, a biweight, a triangular, a tricubic and a rectangular kernel function were applied to extract the salient features of both the long-term (trend) and the short-term (seasonality). The average linear increase growth rates found were mainly attributed to the terrestrial biosphere cycle and changes in the atmospheric circulation regime. The seasonal cycle exhibited a cyclical variation, revealing summer minima for both gases, which may be explained by a biological minimum. Kernel analysis showed two nocturnal CO2 maxima, in spring and autumn, linked to an increase in rainfall. For CO2 daytime records, only the spring peak was detected. As regards CH4, the maximum was located in winter. The best fit for the trend was obtained by the biweight kernel. In contrast, the best adjustment for seasonality was achieved from the Gaussian and the triangular kernel. To sum up, optimal bandwidth selection is important when kernel regression functions are employed. Since no important differences were found between the kernels employed, those which involve least computational effort are recommended.