arXiv subjects

Jun Ma

Publications and source records attributed to Jun Ma.

At least 361 records · Page 20Linked to original sources

arXiv2021

Cutting-edge 3D Medical Image Segmentation Methods in 2020: Are Happy Families All Alike?

Segmentation is one of the most important and popular tasks in medical image analysis, which plays a critical role in disease diagnosis, surgical planning, and prognosis evaluation. During the past five years, on the one hand, thousands of medical image segmentation methods have been proposed for various organs and lesions in different medical images, which become more and more challenging to fairly compare different methods. On the other hand, international segmentation challenges can provide a transparent platform to fairly evaluate and compare different methods. In this paper, we present a comprehensive review of the top methods in ten 3D medical image segmentation challenges during 2020, covering a variety of tasks and datasets. We also identify the "happy-families" practices in the cutting-edge segmentation methods, which are useful for developing powerful segmentation approaches. Finally, we discuss open research problems that should be addressed in the future. We also maintain a list of cutting-edge segmentation methods at \url{https://github.com/JunMa11/SOTA-MedSeg}.

eess.IV↗

arXiv2020

Loss Ensembles for Extremely Imbalanced Segmentation

This short paper briefly presents our methodology details of automatic intracranial aneurysms segmentation from brain MR scans. We use ensembles of multiple models trained from different loss functions. Our method ranked first place in the ADAM challenge segmentation task. The code and trained models are publicly available at https://github.com/JunMa11/ADAM2020.

eess.IV↗

arXiv2020

Exploring Large Context for Cerebral Aneurysm Segmentation

Automated segmentation of aneurysms from 3D CT is important for the diagnosis, monitoring, and treatment planning of the cerebral aneurysm disease. This short paper briefly presents the main technique details of the aneurysm segmentation method in the MICCAI 2020 CADA challenge. The main contribution is that we configure the 3D U-Net with a large patch size, which can obtain the large context. Our method ranked second on the MICCAI 2020 CADA testing dataset with an average Jaccard of 0.7593. Our code and trained models are publicly available at \url{https://github.com/JunMa11/CADA2020}.

eess.IV↗

arXiv2020

Cascaded Framework for Automatic Evaluation of Myocardial Infarction from Delayed-Enhancement Cardiac MRI

Automatic evaluation of myocardium and pathology plays an important role in the quantitative analysis of patients suffering from myocardial infarction. In this paper, we present a cascaded convolutional neural network framework for myocardial infarction segmentation and classification in delayed-enhancement cardiac MRI. Specifically, we first use a 2D U-Net to segment the whole heart, including the left ventricle and the myocardium. Then, we crop the whole heart as a region of interest (ROI). Finally, a new 2D U-Net is used to segment the infraction and no-reflow areas in the whole heart ROI. The segmentation method can be applied to the classification task where the segmentation results with the infraction or no-reflow areas are classified as pathological cases. Our method took second place in the MICCAI 2020 EMIDEC segmentation task with Dice scores of 86.28%, 62.24%, and 77.76% for myocardium, infraction, and no-reflow areas, respectively, and first place in the classification task with an accuracy of 92%.

eess.IV↗

arXiv2020

Combining CNN and Hybrid Active Contours for Head and Neck Tumor Segmentation in CT and PET images

Automatic segmentation of head and neck tumors plays an important role in radiomics analysis. In this short paper, we propose an automatic segmentation method for head and neck tumors from PET and CT images based on the combination of convolutional neural networks (CNNs) and hybrid active contours. Specifically, we first introduce a multi-channel 3D U-Net to segment the tumor with the concatenated PET and CT images. Then, we estimate the segmentation uncertainty by model ensembles and define a segmentation quality score to select the cases with high uncertainties. Finally, we develop a hybrid active contour model to refine the high uncertainty cases. Our method ranked second place in the MICCAI 2020 HECKTOR challenge with average Dice Similarity Coefficient, precision, and recall of 0.752, 0.838, and 0.717, respectively.

eess.IV↗

arXiv2020

Unifying Homophily and Heterophily Network Transformation via Motifs

Higher-order proximity (HOP) is fundamental for most network embedding methods due to its significant effects on the quality of node embedding and performance on downstream network analysis tasks. Most existing HOP definitions are based on either homophily to place close and highly interconnected nodes tightly in embedding space or heterophily to place distant but structurally similar nodes together after embedding. In real-world networks, both can co-exist, and thus considering only one could limit the prediction performance and interpretability. However, there is no general and universal solution that takes both into consideration. In this paper, we propose such a simple yet powerful framework called homophily and heterophliy preserving network transformation (H2NT) to capture HOP that flexibly unifies homophily and heterophily. Specifically, H2NT utilises motif representations to transform a network into a new network with a hybrid assumption via micro-level and macro-level walk paths. H2NT can be used as an enhancer to be integrated with any existing network embedding methods without requiring any changes to latter methods. Because H2NT can sparsify networks with motif structures, it can also improve the computational efficiency of existing network embedding methods when integrated. We conduct experiments on node classification, structural role classification and motif prediction to show the superior prediction performance and computational efficiency over state-of-the-art methods. In particular, DeepWalk-based H2 NT achieves 24% improvement in terms of precision on motif prediction, while reducing 46% computational time compared to the original DeepWalk.

cs.SI↗

arXiv2020

Histogram Matching Augmentation for Domain Adaptation with Application to Multi-Centre, Multi-Vendor and Multi-Disease Cardiac Image Segmentation

Convolutional Neural Networks (CNNs) have achieved high accuracy for cardiac structure segmentation if training cases and testing cases are from the same distribution. However, the performance would be degraded if the testing cases are from a distinct domain (e.g., new MRI scanners, clinical centers). In this paper, we propose a histogram matching (HM) data augmentation method to eliminate the domain gap. Specifically, our method generates new training cases by using HM to transfer the intensity distribution of testing cases to existing training cases. The proposed method is quite simple and can be used in a plug-and-play way in many segmentation tasks. The method is evaluated on MICCAI 2020 M\&Ms challenge, and achieves average Dice scores of 0.9051, 0.8405, and 0.8749, and Hausdorff Distances of 9.996, 12.49, and 12.68 for the left ventricular, myocardium, and right ventricular, respectively. Our results rank the third place in MICCAI 2020 M\&Ms challenge. The code and trained models are publicly available at \url{https://github.com/JunMa11/HM_DataAug}.

eess.IV↗

arXiv2020

EfficientTTS: An Efficient and High-Quality Text-to-Speech Architecture

In this work, we address the Text-to-Speech (TTS) task by proposing a non-autoregressive architecture called EfficientTTS. Unlike the dominant non-autoregressive TTS models, which are trained with the need of external aligners, EfficientTTS optimizes all its parameters with a stable, end-to-end training procedure, while allowing for synthesizing high quality speech in a fast and efficient manner. EfficientTTS is motivated by a new monotonic alignment modeling approach (also introduced in this work), which specifies monotonic constraints to the sequence alignment with almost no increase of computation. By combining EfficientTTS with different feed-forward network structures, we develop a family of TTS models, including both text-to-melspectrogram and text-to-waveform networks. We experimentally show that the proposed models significantly outperform counterpart models such as Tacotron 2 and Glow-TTS in terms of speech quality, training efficiency and synthesis speed, while still producing the speeches of strong robustness and great diversity. In addition, we demonstrate that proposed approach can be easily extended to autoregressive models such as Tacotron 2.

eess.AS↗

arXiv2020

Mixed Information Flow for Cross-domain Sequential Recommendations

Cross-domain sequential recommendation is the task of predict the next item that the user is most likely to interact with based on past sequential behavior from multiple domains. One of the key challenges in cross-domain sequential recommendation is to grasp and transfer the flow of information from multiple domains so as to promote recommendations in all domains. Previous studies have investigated the flow of behavioral information by exploring the connection between items from different domains. The flow of knowledge (i.e., the connection between knowledge from different domains) has so far been neglected. In this paper, we propose a mixed information flow network for cross-domain sequential recommendation to consider both the flow of behavioral information and the flow of knowledge by incorporating a behavior transfer unit and a knowledge transfer unit. The proposed mixed information flow network is able to decide when cross-domain information should be used and, if so, which cross-domain information should be used to enrich the sequence representation according to users' current preferences. Extensive experiments conducted on four e-commerce datasets demonstrate that mixed information flow network is able to further improve recommendation performance in different domains by modeling mixed information flow.

cs.IR↗

arXiv2020

New Determination of Fundamental Properties of Palomar 5 Using Deep DESI Imaging Data

The legacy imaging surveys for the Dark Energy Spectroscopic Instrument project provides multiplecolor photometric data, which are about 2 mag deeper than the SDSS. In this study, we redetermine the fundamental properties for an old halo globular cluster of Palomar 5 based on these new imaging data, including structure parameters, stellar population parameters, and luminosity and mass functions. These characteristics, together with its tidal tails, are key for dynamical studies of the cluster and constraining the mass model of the Milky Way. By fitting the King model to the radial surface density profile of Palomar 5, we derive the core radius of $r_c$ = 2.96' $\pm$ 0.11', tidal radius of $r_t$ = 17.99' $\pm$ 1.49', and concentration parameter of $c$ = 0.78 $\pm$ 0.04. We apply a Bayesian analysis method to derive the stellar population properties and get an age of 11.508 $\pm$ 0.027 Gyr, metallicity of [Fe/H] = -1.798 $\pm$ 0.014, reddening of $E(B-V)$ = 0.0552 $\pm$ 0.0005, and distance modulus of $(m-M)_0$ = 16.835 $\pm$ 0.006. The main-sequence luminosity and mass functions for both the cluster center, and tidal tails are investigated. The luminosity and mass functions at different distances from the cluster center suggest that there is obvious spatial mass segregation. Many faint low-mass stars have been evaporated at the cluster center and the tidal tails are enhanced by low-mass stars. Both the concentration and relaxation times suggest that Palomar 5 is a totally relaxed system.

astro-ph.GA↗

arXiv2020

Towards Data-Efficient Learning: A Benchmark for COVID-19 CT Lung and Infection Segmentation

Purpose: Accurate segmentation of lung and infection in COVID-19 CT scans plays an important role in the quantitative management of patients. Most of the existing studies are based on large and private annotated datasets that are impractical to obtain from a single institution, especially when radiologists are busy fighting the coronavirus disease. Furthermore, it is hard to compare current COVID-19 CT segmentation methods as they are developed on different datasets, trained in different settings, and evaluated with different metrics. Methods: To promote the development of data-efficient deep learning methods, in this paper, we built three benchmarks for lung and infection segmentation based on 70 annotated COVID-19 cases, which contain current active research areas, e.g., few-shot learning, domain generalization, and knowledge transfer. For a fair comparison among different segmentation methods, we also provide standard training, validation and testing splits, evaluation metrics and, the corresponding code. Results: Based on the state-of-the-art network, we provide more than 40 pre-trained baseline models, which not only serve as out-of-the-box segmentation tools but also save computational time for researchers who are interested in COVID-19 lung and infection segmentation. We achieve average Dice Similarity Coefficient (DSC) scores of 97.3\%, 97.7\%, and 67.3\% and average Normalized Surface Dice (NSD) scores of 90.6\%, 91.4\%, and 70.0\% for left lung, right lung, and infection, respectively. Conclusions: To the best of our knowledge, this work presents the first data-efficient learning benchmark for medical image segmentation and the largest number of pre-trained models up to now. All these resources are publicly available, and our work lays the foundation for promoting the development of deep learning methods for efficient COVID-19 CT segmentation with limited data.

eess.IV↗

arXiv2020

Estimating ages and metallicities of M31 star clusters from LAMOST DR6

Context. Determining the metallicities and ages of M31 clusters is fundamental to the study of the formation and evolution of M31 itself. The Large Sky Area Multi-Object Fiber Spectroscopic Telescope (LAMOST) has carried out a systematic spectroscopic campaign of clusters and candidates in M31. Aims. We constructed a catalogue of 346 M31 clusters observed by LAMOST. By combining the information of the LAMOST spectra and the multi-band photometry, we developed a new algorithm to estimate the metallicities and ages of these clusters. Methods. We distinguish young clusters from old using random forest classifiers based on a empirical training data set selected from the literature. Ages of young clusters are derived from the spectral energy distribution (SED) fits of their multi-band photometric measurements. Their metallicities are estimated by fitting their observed spectral principal components extracted from the LAMOST spectra with those from the young metal-rich single stellar population (SSP) models. For old clusters, we built non-parameter random forest models between the spectral principal components and/or multi-band colours and the parameters of the clusters based on a training data set constructed from the SSP models. The ages and metallicities of the old clusters are then estimated by fitting their observed spectral principal components extracted from the LAMOST spectra and multi-band colours from the photometric measurements with the resultant random forest models. Results. We derived parameters of 53 young and 293 old clusters in our catalogue. Our resultant parameters are in good agreement with those from the literature. The ages of about 30 catalogued clusters and metallicities of about 40 sources are derived for the first time.

astro-ph.GA↗

arXiv2020

Two substructures in the nearby stellar halo found in Gaia and RAVE

We use the second Gaia data release (Gaia DR2), combined with RAVE spectroscopic surveys, to identify the substructures in the nearby stellar halo. We select 3,845 halo stars kinematically and chemically, and determine their density distribution in energy and angular momentum space. To select the substructures from overdensities, we reshuffle the velocities and estimate their significance. Two statistically significant substructures, GR-1 and GR-2, are identified. GR-1 has a high binding energy and small $z$-angular momentum. GR-2 is metal-rich but retrograde. They are both new substructure, and may be accretion debris of dwarf galaxies.

astro-ph.GA↗

arXiv2020

Existence of the Metal-Rich Stellar Halo and High-velocity Thick Disk in the Galaxy

Based on the second Gaia data release (DR2), combined with the LAMOST and APOGEE spectroscopic surveys, we study the kinematics and metallicity distribution of the high-velocity stars that have a relative speed of at least 220 ${\rm km\ s^{-1}}$ with respect to the local standard of rest in the Galaxy. The rotational velocity distribution of the high-velocity stars with [Fe/H] $>-1.0$ dex can be well described by a two-Gaussian model, with peaks at $V_ϕ\sim +164.2\pm0.7$ and $V_ϕ\sim +3.0\pm0.3$ ${\rm km\ s^{-1}}$, associated with the thick disk and halo, respectively. This implies that there should exist a high-velocity thick disk (HVTD) and a metal-rich stellar halo (MRSH) in the Galaxy. The HVTD stars have the same position as the halo in the Toomre diagram and but show the same rotational velocity and metallicity as the canonical thick disk. The MRSH stars have basically the same rotational velocity, orbital eccentricity, and position in the Lindblad and Toomre diagram as the canonical halo stars, but they are more metal-rich. Furthermore, the metallicity distribution function (MDF) of our sample stars are well fitted by a four-Gaussian model, associated with the outer-halo, inner-halo, MRSH, and HVTD, respectively. Chemical and kinematic properties and age imply that the MRSH and HVTD stars may form in situ.

astro-ph.GA↗

arXiv2020

Eulerian pairs and Eulerian recurrence systems

In this paper, we characterize a duality relation between Eulerian recurrences and Eulerian recurrence systems, which generalizes and unifies Hermite-Biehler decompositions of several enumerative polynomials, including flag descent polynomials for hyperoctahedral group, flag ascent-plateau polynomials for Stirling permutations, up-down run polynomials for symmetric group and alternating run polynomials for hyperoctahedral group. As applications, we derive some properties of associated enumerative polynomials. In particular, we find that both the ascent-plateau polynomials and left ascent-plateau polynomials for Stirling permutations are alternatingly increasing, and so they are unimodal with modes in the middle.

math.CO↗

arXiv2020

Peak thermoelectric power factor of holey silicon films

The thermoelectric properties of nanostructured silicon are not fully understood despite their initial promise. While the anomalously low thermal conductivity has attracted much work, the impact of nanostructuring on the power factor has mostly escaped attention. While initial reports did not find any significant changes to the power factor compared to the bulk, subsequent detailed measurements on p-type silicon nanowires showed a stark reduction in the Seebeck coefficient when compared to similarly doped bulk. The reduction is consistent with the disappearance of the phonon drag contribution, due to phonon boundary scattering. Here, we report measurements on a different nanostructure, holey silicon films, to test if similar loss of phonon drag can be observed. By devising experiments where all properties are measured on the same sample, we show that though these films possess electrical conductivity close to that in the bulk at comparable doping, they exhibit considerably smaller thermopower. The data are consistent with loss of phonon drag. At neck distances between 120 - 230 nm, the power factor at optimal doping is $\sim$ 50 percent that of the bulk. These insights are useful in the practical design of future thermoelectric devices based on nanostructured silicon.

cond-mat.mtrl-sci↗

arXiv2020

On Robust Stability and Performance with a Fixed-Order Controller Design for Uncertain Systems

Typically, it is desirable to design a control system that is not only robustly stable in the presence of parametric uncertainties but also guarantees an adequate level of system performance. However, most of the existing methods need to take all extreme models over an uncertain domain into consideration, which then results in costly computation. Also, since these approaches attempt (rather unrealistically) to guarantee the system performance over a full frequency range, a conservative design is always admitted. Here, taking a specific viewpoint of robust stability and performance under a stated restricted frequency range (which is applicable in rather many real-world situations), this paper provides an essential basis for the design of a fixed-order controller for a system with bounded parametric uncertainties. A Hurwitz polynomial is used in the design and the robust stability is characterized by the notion of positive realness, such that the required robust stability condition is then suitably successfully constructed. Also, the robust performance criteria in terms of sensitivity shaping under different frequency ranges are constructed based on an approach of bounded realness analysis. Necessary and sufficient conditions are provided for both the robust stability and robust performance criteria. Furthermore, these conditions are expressed in the framework of linear matrix inequality (LMI) constraints, and thus can be efficiently solved. Comparative simulations are provided to illustrate the effectiveness and efficiency of the proposed approach.

math.OC↗

arXiv2020

Grasping Detection Network with Uncertainty Estimation for Confidence-Driven Semi-Supervised Domain Adaptation

Data-efficient domain adaptation with only a few labelled data is desired for many robotic applications, e.g., in grasping detection, the inference skill learned from a grasping dataset is not universal enough to directly apply on various other daily/industrial applications. This paper presents an approach enabling the easy domain adaptation through a novel grasping detection network with confidence-driven semi-supervised learning, where these two components deeply interact with each other. The proposed grasping detection network specially provides a prediction uncertainty estimation mechanism by leveraging on Feature Pyramid Network (FPN), and the mean-teacher semi-supervised learning utilizes such uncertainty information to emphasizing the consistency loss only for those unlabelled data with high confidence, which we referred it as the confidence-driven mean teacher. This approach largely prevents the student model to learn the incorrect/harmful information from the consistency loss, which speeds up the learning progress and improves the model accuracy. Our results show that the proposed network can achieve high success rate on the Cornell grasping dataset, and for domain adaptation with very limited data, the confidence-driven mean teacher outperforms the original mean teacher and direct training by more than 10% in evaluation loss especially for avoiding the overfitting and model diverging.

cs.CV↗