SPARC Dataset Research Articles

Bounding the cosmological constant using galactic rotation curves from SPARC dataset

Bounding the cosmological constant using galactic rotation curves from SPARC dataset

An image registration-based self-supervised Su-Net for carotid plaque ultrasound image segmentation

Background and ObjectivesTotal Plaque Area (TPA) measurement is critical for early diagnosis and intervention of carotid atherosclerosis in individuals with high risk for stroke. The delineation of the carotid plaques is necessary for TPA measurement, and deep learning methods can automatically segment the plaque and measure TPA from carotid ultrasound images. A large number of labeled images is essential for training a good deep learning model, but it is very difficult to collect such large labeled datasets for carotid image segmentation in clinical practice. Self-supervised learning can provide a possible solution to improve the deep-learning models on small labeled training datasets by designing a pretext task to pre-train the models without using the segmentation masks. However, the existing self-supervised learning methods do not consider the feature presentations of object contours. MethodsIn this paper, we propose an image registration-based self-supervised learning method and a stacked U-Net (SSL-SU-Net) for carotid plaque ultrasound image segmentation, which can better exploit the semantic features of carotid plaque contours in self-supervised task training. ResultsOur network was trained on different numbers of labeled images (n = 10, 33, 50 and 100 subjects) and tested on 44 subjects from the SPARC dataset (n = 144, London, Canada). The network trained on the entire SPARC dataset was then directly applied to an independent dataset collected in Zhongnan hospital (n = 497, Wuhan, China). For the 44 subjects tested on the SPARC dataset, our method yielded a DSC of 80.25-89.18% and the produced TPA measurements, which were strongly correlated with manual segmentation (r = 0.965-0.995, ρ< 0.0001). For the Zhongnan dataset, the DSC was 90.3% and algorithm TPAs were strongly correlated with manual TPAs (r = 0.985, ρ< 0.0001). ConclusionsThe results demonstrate that our proposed method yielded excellent performance and good generalization ability when trained on a small labeled dataset, facilitating the use of deep learning in carotid ultrasound image analysis and clinical practice. The code of our algorithm is available https://github.com/a610lab/Registration-SSL.

Deep Learning-Based Measurement of Total Plaque Area in B-Mode Ultrasound Images.

Measurement of total-plaque-area (TPA) is important for determining long term risk for stroke and monitoring carotid plaque progression. Since delineation of carotid plaques is required, a deep learning method can provide automatic plaque segmentations and TPA measurements; however, it requires large datasets and manual annotations for training with unknown performance on new datasets. A UNet++ ensemble algorithm was proposed to segment plaques from 2D carotid ultrasound images, trained on three small datasets (n=33, 33, 34 subjects) and tested on 44 subjects from the SPARC dataset (n=144, London, Canada). The ensemble was also trained on the entire SPARC dataset and tested with a different dataset (n=497, Zhongnan Hospital, China). Algorithm and manual segmentations were compared using Dice-similarity-coefficient (DSC), and TPAs were compared using the difference ( ∆TPA), Pearson correlation coefficient (r) and Bland-Altman analyses. Segmentation variability was determined using the intra-class correlation coefficient (ICC) and coefficient-of-variation (CoV). For 44 SPARC subjects, algorithm DSC was 83.3-85.7%, and algorithm TPAs were strongly correlated (r=0.985-0.988; p<0.001) with manual results with marginal biases (0.73-6.75)mm 2 using the three training datasets. Algorithm ICC for TPAs (ICC=0.996) was similar to intra- and inter-observer manual results (ICC=0.977, 0.995). Algorithm CoV=6.98% for plaque areas was smaller than the inter-observer manual CoV (7.54%). For the Zhongnan dataset, DSC was 88.6% algorithm and manual TPAs were strongly correlated (r=0.972, p<0.001) with ∆TPA=-0.44 ±4.05mm 2 and ICC=0.985. The proposed algorithm trained on small datasets and segmented a different dataset without retraining with accuracy and precision that may be useful clinically and for research.

A Multiple-Integration Encoder for Multi-Turn Text-to-SQL Semantic Parsing

This paper studies multi-turn text-to-SQL generation, which is a new but important task in semantic parsing. In order to deal with its two challenges, i.e., multi-turn interaction and cross-domain evaluation, this paper proposes a multiple-integration encoder, which derives the vector representations of user utterances and database schemas using three custom-designed modules for information integration. First, an utterance representation enhancing module is built to integrate the information of history utterances into the representation of each token in current utterance by attentive selection. Second, a schema discrepancy enhancing module is designed to integrate previous predicted SQL query into the representation of schema items. Third, a latent schema linking module is employed to integrate schema information into utterance representations for better dealing with unseen database schemas. These three modules are all implemented based on a lightweight multi-head attention mechanism, which reduces the number of parameters in conventional multi-head attention. Experimental results on the SParC dataset show that our method achieved better accuracy of multi-turn text-to-SQL generation than the most advanced benchmarks. Further ablations studies and analysis also demonstrate the effectiveness of the three modules designed for information integration in the encoder.

SPARC: SODA, an interactive software for curating SPARC datasets

In this era of data‐intensive research, data sharing and reuse are becoming increasingly important for rapid discovery and innovation. While sharing can be as easy as a few clicks thanks to numerous data sharing platforms, data reuse is still an issue as standards for data curation and annotation are not rigorously or consistently followed by all researchers. As a result, it is often difficult, if not impossible, for researchers to understand and work with data they were not involved with during collection or analysis. To address this concern, the NIH SPARC program has imposed strict guidelines for curating and sharing data resulting from SPARC‐funded research. Complying with the guidelines requires additional time investment from the researchers (as with any data curation and submission standards), and the data curation process could thus become overwhelming as more and more data is generated. The objective of this project was to develop computer software to improve the curation workflow for SPARC researchers.To achieve that, we are developing a software named SODA (Software for Organizing Data Automatically) to assist SPARC investigators in curating and annotating their datasets. Distributed as an open‐source desktop application (Windows, MAC, Linux), the goal of SODA is to bridge a long‐standing and overlooked gap between comprehensive data standards and their convenient application by researchers. SODA provides an interactive interface that, without requiring any coding knowledge, walks SPARC investigators step‐by‐step through the data organization and sharing requirements, while automating repetitive, complex and/or time‐consuming tasks (Fig 1). An introduction to the software and the algorithms implemented will be given. Quantitative results from ongoing performance evaluation (time to curate a SPARC dataset with and without SODA) by beta testers, all of whom receive funding from the SPARC program, will be presented as well.Overall, the beta testers for early versions have acknowledged that SODA has the potential to improve their curation workflow. Data curation is becoming a critical issue, even beyond the SPARC consortium, and SODA could impact the broader research community by providing an exemplar and foundational tool for convenient and time‐efficient curation of data, which could be adopted by other projects. Potential applications, beyond the SPARC program, will be thus discussed as well.Support or Funding InformationSPARC OT2OD025308‐01S2Overview of the major features included in SODA during the on‐going development phaseFigure 1

An Alternative Model of Rotation Curve that Explains Anomalous Orbital Velocity, Mass Discrepancy and Structure of Some Galaxies

A new model of galaxy rotation based on the cyclostrophic model of vortices found in nature is developed. The model is tested using the SPARC dataset of 175 galaxies and a smaller dataset comprising of 60 galaxies. Analysis of the datasets showed that galactic rotation can be adequately described using the observed surface brightness of galaxies and the newly developed cyclostrophic velocity model. The use of the luminosity and the inverse mass-to-light ratio in lieu of the surface brightness, also yield a very good fit of the observed and computed galaxy rotation velocity. Evidently, galactic rotation greatly depends on the cyclostrophic balance of the pressure gradient and the centrifugal forces and the seismic-induced radial expansion occurring in various stars. This is the most probable origin of the action of a single force law that has been overlooked in previous studies. Therefore, the need for a super-massive black hole at the center of galaxies or hidden dark matter can be eliminated. Attractive gravitational force can occur even without a massive black hole at the center of galaxies. There appears to be a pressure gradient force between the center and the outer parts of galaxies that sustains attraction. The cyclostrophic model appears to be the physical basis of the Tully-Fisher relation. Furthermore, the missing mass problem associated with galactic rotation can be attributed to the orbital expansion of celestial objects perturbed by seismic-induced forces. In addition, massive tremors or starquakes may create a domino effect in perturbing nearby stars along the axis of the seismic-induced force and this could result in the formation of elliptical galaxies as the orbits of seismic-perturbed neighboring stars become larger.

Searching for a possible dipole anisotropy in acceleration scale with 147 rotationally supported galaxies**Supported by National Natural Science Foundation of China (11675182, 11690022, 11603005)

We report a possible dipole anisotropy in acceleration scale g† with 147 rotationally supported galaxies in the local Universe. It is found that a monopole and dipole correction for the radial acceleration relation can better describe the SPARC data set. The monopole term is negligible but the dipole magnitude is significant. It is also found that the dipole correction is mostly induced by anisotropy in the acceleration scale. The magnitude of the -dipole reaches 0.25 ± 0.04, and its direction is aligned to (l,b) = (171.30° ± 7.18°, −15.41° ± 4.87°), which is very close to the maximum anisotropy direction from the hemisphere comparison method. Furthermore, a robust check shows that the dipole anisotropy could not be reproduced by an isotropic mock data set. However, it is still premature to claim that the Universe is anisotropic, due to the small data samples and uncertainty in the current observations.

Stellar feedback and the energy budget of late-type Galaxies: missing baryons and core creation

In a $\Lambda$CDM cosmology, galaxy formation is a globally inefficient process: it is often the case that far fewer baryons are observed in galaxy disks than expected from the cosmic baryon fraction. The location of these "missing baryons" is unclear. By fitting halo profiles to the rotation curves of galaxies in the SPARC data set, we measure the "missing baryon" mass for individual late-type systems. Assuming that haloes initially accrete the cosmological baryon fraction, we show that the maximum energy available from supernovae is typically not enough to completely eject these "missing baryons" from a halo, but it is often sufficient to heat them to the virial temperature. The energy available from supernovae has the same scaling with galaxy mass as the energy needed to heat or eject the "missing baryons", indicating that the coupling efficiency of the feedback to the ISM may be constant with galaxy virial mass. We further find that the energy available from supernova feedback is always enough to convert a primordial cusp into a core and has magnitude consistent with what is required to heat the "missing baryons" to the virial temperature. Taking a census of the baryon content of galaxies with ${\rm 10^9<M_{vir}/M_{\odot}<10^{12}}$ reveals that $\sim86\%$ of baryons are likely to be in a hot phase surrounding the galaxies and possibly observable in the X-ray, $\sim7\%$ are in the form of cold gas, and $\sim7\%$ are in stars.

Multistate scalar field dark matter and its correlation with galactic properties

In this paper, we search for the correlations between the intrinsic properties of galaxies and the Bose–Einstein condensate (BEC) under the scheme of a scalar field dark matter (SFDM) at the temperature of condensation greater than zero. According to this paradigm the BEC is distributed in several states. Based on the galactic rotation curves collected in SPARC dataset, we observe that the SFDM parameters present a weak correlation with the most of the galaxy properties, having only a correlation with those related to neutral hydrogen emissions. In addition, we found evidence to the support of self-interaction between the different BEC states proposing that, in future studies, the crossed terms in SFDM equations must be considered. Finally, we find a null correlation with galaxy distances giving support to the nonhierarchy of SFDM formation.

NIHAO – XIV. Reproducing the observed diversity of dwarf galaxy rotation curve shapes in ΛCDM

The significant diversity of rotation curve (RC) shapes in dwarf galaxies has recently emerged as a challenge to LCDM: in dark matter (DM) only simulations, DM halos have a universal cuspy density profile that results in self-similar RC shapes. We compare RC shapes of simulated galaxies from the NIHAO project with observed galaxies from the homogeneous SPARC dataset. The DM halos of the NIHAO galaxies can expand to form cores, with the degree of expansion depending on their stellar-to-halo mass ratio. By means of the $V_{\rm 2kpc}-V_{\rm Rlast}$ relation (where $V_{\rm Rlast}$ is the outermost measured rotation velocity), we show that both the average trend and the scatter in RC shapes of NIHAO galaxies are in reasonable agreement with SPARC: this represents a significant improvement compared to simulations that do not result in DM core formation, suggesting that halo expansion is a key process in matching the diversity of dwarf galaxy RCs. Note that NIHAO galaxies can reproduce even the extremely slowly rising RCs of IC 2574 and UGC 5750. Revealingly, the range where observed galaxies show the highest diversity corresponds to the range where core formation is most efficient in NIHAO simulations, 50$<V_{\rm Rlast}$\km s$^{-1}<$100. A few observed galaxies in this range cannot be matched by any NIHAO RC nor by simulations that predict a universal halo profile. Interestingly, the majority of these are starbursts or emission-line galaxies, with steep RCs and small effective radii. Such galaxies represent an interesting observational target providing new clues to the process/viability of cusp-core transformation, the relationship between starburst and inner potential well, and the nature of DM.

Galaxy rotations from quantised inertia and visible matter only

It is shown here that a model for inertial mass, called quantised inertia, or MiHsC (Modified inertia by a Hubble-scale Casimir effect) predicts the rotational acceleration of the 153 good quality galaxies in the SPARC dataset (2016 AJ 152 157), with a large range of scales and mass, from just their visible baryonic matter, the speed of light and the co-moving diameter of the observable universe. No dark matter is needed. The performance of quantised inertia is comparable to that of MoND, yet it needs no adjustable parameter. As a further critical test, quantised inertia uniquely predicts a specific increase in the galaxy rotation anomaly at higher redshifts. This test is now becoming possible and new data shows that galaxy rotational accelerations do increase with redshift in the predicted manner, at least up to Z=2.2.

Toward a Semiautomatic Machine Learning Retrieval of Biophysical Parameters

Biophysical parameters such as leaf chlorophyll content (LCC) and leaf area index (LAI) are standard vegetation products that can be retrieved from Earth observation imagery. This paper introduces a new machine learning regression algorithms (MLRAs) toolbox into the scientific Automated Radiative Transfer Models Operator (ARTMO) software package. ARTMO facilitates retrieval of biophysical parameters from remote observations in a MATLAB graphical user interface (GUI) environment. The MLRA toolbox enables analyzing the predictive power of various MLRAs in a semiautomatic and systematic manner, and applying a selected MLRA to multispectral or hyperspectral imagery for mapping applications. It contains both linear and nonlinear state-of-the-art regression algorithms, in particular linear feature extraction via principal component regression (PCR), partial least squares regression (PLSR), decision trees (DTs), neural networks (NNs), kernel ridge regression (KRR), and Gaussian processes regression (GPR). The performance of multiple implemented regression strategies has been evaluated against the SPARC dataset (Barrax, Spain) and simulated Sentinel-2 (8 bands), CHRIS (62 bands) and HyMap (125 bands) observations. In general, nonlinear regression algorithms (NN, KRR, and GPR) outperformed linear techniques (PCR and PLSR) in terms of accuracy, bias, and robustness. Most robust results along gradients of training/validation partitioning and noise variance were obtained by KRR while GPR delivered most accurate estimations. We applied a GPR model to a hyperspectral HyMap flightline to map LCC and LAI. We exploited the associated uncertainty intervals to gain insight in the per-pixel performance of the model.

Gaussian Process Retrieval of Chlorophyll Content From Imaging Spectroscopy Data

Precise and spatially-explicit knowledge of leaf chlorophyll content <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex Notation="TeX">$(Chl)$</tex></formula> is crucial to adequately interpret the chlorophyll fluorescence <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex Notation="TeX">$(ChF)$</tex></formula> signal from space. Accompanying information about the reliability of the <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex Notation="TeX">$Chl$</tex></formula> estimation becomes more important than ever. Recently, a new statistical method was proposed within the family of nonparametric Bayesian statistics, namely Gaussian Processes regression (GPR). GPR is simpler and more robust than their machine learning family members while maintaining very good numerical performance and stability. Other features include: i) GPR requires a relatively small training data set and can adopt very flexible kernels, ii) GPR identifies the relevant bands and observations in establishing relationships with a variable, and finally iii) along with pixelwise estimations GPR provides accompanying confidence intervals. We used GPR to retrieve <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex Notation="TeX">$Chl$</tex> </formula> from hyperspectral reflectance data and evaluated the portability of the regression model to other images. Based on field <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex Notation="TeX">$Chl$</tex></formula> measurements from the SPARC dataset and corresponding spaceborne CHRIS spectra (acquired in 2003, Barrax, Spain), GPR developed a regression model that was excellently validated ( <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex Notation="TeX">$r^{2}$</tex></formula> : 0.96, RMSE: 3.82 <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex Notation="TeX">$\mu{\rm g/cm}^{2}$</tex> </formula> ). The SPARC-trained GPR model was subsequently applied to CHRIS images (Barrax, 2003, 2009) and airborne CASI flightlines (Barrax 2009) to generate <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex Notation="TeX">$Chl$</tex></formula> maps. The accompanying confidence maps provided insight in the robustness of the retrievals. Similar confidences were achieved by both sensors, which is encouraging for upscaling <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex Notation="TeX">$Chl$</tex> </formula> estimates from field to landscape scale. Because of its robustness and ability to deliver confidence intervals, GPR is evaluated as a promising candidate for implementation into <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex Notation="TeX">$ChF$</tex> </formula> processing chains.