arXiv subjects

Yiming Li

Publications and source records attributed to Yiming Li.

323 records · Page 18Linked to original sources

arXiv2019

Boundary Effect-Aware Visual Tracking for UAV with Online Enhanced Background Learning and Multi-Frame Consensus Verification

Due to implicitly introduced periodic shifting of limited searching area, visual object tracking using correlation filters often has to confront undesired boundary effect. As boundary effect severely degrade the quality of object model, it has made it a challenging task for unmanned aerial vehicles (UAV) to perform robust and accurate object following. Traditional hand-crafted features are also not precise and robust enough to describe the object in the viewing point of UAV. In this work, a novel tracker with online enhanced background learning is specifically proposed to tackle boundary effects. Real background samples are densely extracted to learn as well as update correlation filters. Spatial penalization is introduced to offset the noise introduced by exceedingly more background information so that a more accurate appearance model can be established. Meanwhile, convolutional features are extracted to provide a more comprehensive representation of the object. In order to mitigate changes of objects' appearances, multi-frame technique is applied to learn an ideal response map and verify the generated one in each frame. Exhaustive experiments were conducted on 100 challenging UAV image sequences and the proposed tracker has achieved state-of-the-art performance.

cs.CV↗

arXiv2019

Learning Aberrance Repressed Correlation Filters for Real-Time UAV Tracking

Traditional framework of discriminative correlation filters (DCF) is often subject to undesired boundary effects. Several approaches to enlarge search regions have been already proposed in the past years to make up for this shortcoming. However, with excessive background information, more background noises are also introduced and the discriminative filter is prone to learn from the ambiance rather than the object. This situation, along with appearance changes of objects caused by full/partial occlusion, illumination variation, and other reasons has made it more likely to have aberrances in the detection process, which could substantially degrade the credibility of its result. Therefore, in this work, a novel approach to repress the aberrances happening during the detection process is proposed, i.e., aberrance repressed correlation filter (ARCF). By enforcing restriction to the rate of alteration in response maps generated in the detection phase, the ARCF tracker can evidently suppress aberrances and is thus more robust and accurate to track objects. Considerable experiments are conducted on different UAV datasets to perform object tracking from an aerial view, i.e., UAV123, UAVDT, and DTB70, with 243 challenging image sequences containing over 90K frames to verify the performance of the ARCF tracker and it has proven itself to have outperformed other 20 state-of-the-art trackers based on DCF and deep-based frameworks with sufficient speed for real-time applications.

cs.CV↗

arXiv2019

Clearly Discriminate the Continuum Band and Exciton State of the Hybrid Lead Bromide Perovskite

Electronic states of the hybrid perovskite enable their promising applications as distinctive optoelectronic materials. The understanding of their electronic structures and charge characters remains highly controversial. The electronic mechanism such as reabsorption, Urbach tail and indirect band for interpreting dual-peak emissions is one of the controversial focuses. Herein, we report that through heterojunction enhanced exciton dissociation and global tracing of multiple radiative electronic states across wide temperature regions, we have succeeded in directly observing free carrier emissions from the hybrid lead bromide perovskite and clearly discriminating the direct continuum band and exciton states. The widely-concerned dual-peak emissions are clarified to be excitonic, arising from two types of exciton states of the perovskite. These excitons possess giant binding energies and superior phase stability compared to conventional inorganic semiconductors, providing important implications for exploiting the excitonic mechanism for realizing novel optoelectronic applications.

physics.app-ph↗

arXiv2019

Semi-Supervised Brain Lesion Segmentation with an Adapted Mean Teacher Model

Automated brain lesion segmentation provides valuable information for the analysis and intervention of patients. In particular, methods based on convolutional neural networks (CNNs) have achieved state-of-the-art segmentation performance. However, CNNs usually require a decent amount of annotated data, which may be costly and time-consuming to obtain. Since unannotated data is generally abundant, it is desirable to use unannotated data to improve the segmentation performance for CNNs when limited annotated data is available. In this work, we propose a semi-supervised learning (SSL) approach to brain lesion segmentation, where unannotated data is incorporated into the training of CNNs. We adapt the mean teacher model, which is originally developed for SSL-based image classification, for brain lesion segmentation. Assuming that the network should produce consistent outputs for similar inputs, a loss of segmentation consistency is designed and integrated into a self-ensembling framework. Specifically, we build a student model and a teacher model, which share the same CNN architecture for segmentation. The student and teacher models are updated alternately. At each step, the student model learns from the teacher model by minimizing the weighted sum of the segmentation loss computed from annotated data and the segmentation consistency loss between the teacher and student models computed from unannotated data. Then, the teacher model is updated by combining the updated student model with the historical information of teacher models using an exponential moving average strategy. For demonstration, the proposed approach was evaluated on ischemic stroke lesion segmentation, where it improves stroke lesion segmentation with the incorporation of unannotated data.

cs.CV↗

arXiv2019

Low-threshold Stimulated Emission of Hybrid Perovskites at Room Temperature through Defect-Mediated Bound Excitons

Excitonic stimulated emission provides a promising mechanism and route to achieve low-threshold semiconductor lasers for micro-nano optoelectronic integrations. However, excitonic stimulated emission from quantum structure-free semiconductors has rarely been realised at room temperature due to the phase transition between excitonic and electron-hole plasma states. Herein, we show that through trap-state and band-edge engineering, bound exciton states can be stabilised within the hybrid lead bromide perovskite. Under modest pumping conditions, these states enable stimulated emission behaviour that exhibits a low threshold carrier density of only 1.6*1017 cm-3, as well as a high peak gain coefficient of ~1300 cm-1. This is the first time that bound exciton stimulated emission has been realised at room temperature from a quantum structure-free semiconductor. Not only does this open up new research horizons for perovskite materials, but also it has important implications for semiconductor excitonic physics and the development of next-generation optoelectronic applications.

physics.app-ph↗

arXiv2017

Identification of Free and Bound Exciton States and Their Phase-Dependent Trapping Behavior in Lead Halide Perovskites

In this work we probe the sub-gap energy states within polycrystalline and single crystal lead halide perovskites to better understand their intrinsic photophysics behaviors. Through combined temperature and intensity-dependent optical measurements, we reveal the existence of both free and bound exciton contributions within the sub-gap energy state manifold. The trapping and recombination dynamics of these excitons is shown to be strongly dependent on the structural phase of the perovskite. The orthorhombic phase exhibits ultrafast exciton trapping and distinct trap emission, while the tetragonal phase gives low monomolecular recombination velocity and capture cross-sections (~10-18 cm2). Within the multiphonon transition scenario, this suppression in charge trapping is caused by the increase in the charge capture activation energy due to the reduction in electron-lattice interactions, which can be the origin for the unexpected long carrier lifetime in these material systems.

cond-mat.mtrl-sci↗

arXiv2015

Point source detection performance of Hard X-ray Modulation Telescope imaging observation

The Hard X-ray Modulation Telescope (HXMT) will perform an all-sky survey in hard X-ray band as well as deep imaging of a series of small sky regions. We expect various compact objects to be detected in these imaging observations. Point source detection performance of HXMT imaging observation depends not only on the instrument but also on its data analysis since images are reconstructed from HXMT observed data with numeric methods. Denoising technique plays an import part in HXMT imaging data analysis pipeline alongside with demodulation and source detection. In this paper we have implemented several methods for denoising HXMT data and evaluated the point source detection performances in terms of sensitivities and location accuracies. The results show that direct demodulation with 1-fold cross correlation should be the default reconstruction and regularization methods, although both sensitivity and location accuracy could be further imporved by selecting and tuning numerical methods in data analysis of HXMT imaging observations.

astro-ph.IM↗

arXiv2014

Quarkonia production in proton-lead collisions at LHCb

The production of $J/ψ$ and $Υ$ mesons decaying into dimuon final state is studied at the LHCb experiment, with rapidity in the range of (1.5, 4.0) or (-5.0, -2.5) in proton-lead collisions at a nucleon-nucleon centre-of-mass energy $\sqrt{s_{NN}}$ = 5 TeV, based on a data sample corresponding to an integrated luminosity of 1.6 /fb. The nuclear modification factor and forward-backward production ratio are determined for prompt $J/ψ$, $J/ψ$ from b-hadron decay and $Υ(1S)$ mesons in study of the cold nuclear matter effects.

hep-ex↗

arXiv2014

Heavy flavour spectroscopy at LHC

The pp collision data collected in the LHC Run I provides a great opportunity for heavy flavour studies. The latest results on exotic states, heavy baryon and $B_c^+$ mesons are reviewed.

hep-ex↗

arXiv2014

Software-Defined Networking: State of the Art and Research Challenges

Plug-and-play information technology (IT) infrastructure has been expanding very rapidly in recent years. With the advent of cloud computing, many ecosystem and business paradigms are encountering potential changes and may be able to eliminate their IT infrastructure maintenance processes. Real-time performance and high availability requirements have induced telecom networks to adopt the new concepts of the cloud model: software-defined networking (SDN) and network function virtualization (NFV). NFV introduces and deploys new network functions in an open and standardized IT environment, while SDN aims to transform the way networks function. SDN and NFV are complementary technologies; they do not depend on each other. However, both concepts can be merged and have the potential to mitigate the challenges of legacy networks. In this paper, our aim is to describe the benefits of using SDN in a multitude of environments such as in data centers, data center networks, and Network as Service offerings. We also present the various challenges facing SDN, from scalability to reliability and security concerns, and discuss existing solutions to these challenges.

cs.NI↗

arXiv2013

Star Formation Rates in Resolved Galaxies: Calibrations with Near and Far Infrared Data for NGC5055 and NGC6946

We use the near--infrared Brγhydrogen recombination line as a reference star formation rate (SFR) indicator to test the validity and establish the calibration of the {\it Herschel} PACS 70 μm emission as a SFR tracer for sub--galactic regions in external galaxies. Brγoffers the double advantage of directly tracing ionizing photons and of being relatively insensitive to the effects of dust attenuation. For our first experiment, we use archival CFHT Brγand Ks images of two nearby galaxies: NGC\,5055 and NGC\,6946, which are also part of the {\it Herschel} program KINGFISH (Key Insights on Nearby Galaxies: a Far-Infrared Survey with Herschel). We use the extinction corrected Brγemission to derive the SFR(70) calibration for H{\sc ii} regions in these two galaxies. A comparison of the SFR(70) calibrations at different spatial scales, from 200 pc to the size of the whole galaxy, reveals that about 50% of the total 70μm emission is due to dust heated by stellar populations that are unrelated to the current star formation. We use a simple model to qualitatively relate the increase of the SFR(70) calibration coefficient with decreasing region size to the star formation timescale. We provide a calibration for an unbiased SFR indicator that combines the observed Hαwith the 70 μm emission, also for use in H{\sc ii} regions. We briefly analyze the PACS 100 and 160 μm maps and find that longer wavelengths are not as good SFR indicators as 70μm, in agreement with previous results. We find that the calibrations show about 50% difference between the two galaxies, possibly due to effects of inclination.

astro-ph.CO↗

arXiv2010

Spitzer 70~$μ$m Emission as a SFR Indicator for Sub--Galactic Regions

We use Spitzer 24 $μ$m, 70 $μ$m and ground based H$α$ data for a sample of 40 SINGS galaxies to establish a star formation rate (SFR) indicator using 70 $μ$m emission for sub--galactic ($\sim0.05-2\ \rm{kpc}$) line-emitting regions and to investigate limits in application. A linear correlation between 70 $μ$m and SFR is found and a star formation indicator SFR(70) is proposed for line-emitting sub-galactic regions as $\rm Σ(SFR)\ ({M_{\odot}\cdot yr^{-1}\cdot kpc^{-2}})=9.4\times10^{-44}\ Σ(70)\ \rm{(ergs\cdot s^{-1}\cdot kpc^{-2})}$, for regions with $12+\rm{log(O/H)}\gtrsim8.4$ and $\rm Σ(SFR)\gtrsim10^{-3}\ (M_{\odot}\cdot yr^{-1}\cdot kpc^{-2})$, with a 1-$σ$ dispersion around the calibration of $\sim0.16$ dex. We also discuss the influence of metallicity on the scatter of the data. Comparing with the SFR indicator at 70 $μ$m for integrated light from galaxies, we find that there is $\sim40%$ excess 70 $μ$m emission in galaxies, which can be attributed to stellar populations not involved in the current star formation activity.

astro-ph.CO↗

arXiv2010

ISIS2: Pixel Sensor with Local Charge Storage for ILC Vertex Detector

ISIS (In-situ Storage Imaging Sensor) is a novel CMOS sensor with multiple charge storage capability developed for the ILC vertex detector by the Linear Collider Flavour Identification (LCFI) collaboration. This paper reports test results for ISIS2, the second generation of ISIS sensors implemented in a 0.18 micron CMOS process. The local charge storage and charge transfer were unambiguously demonstrated.

physics.ins-det↗

arXiv2010

Chargino and Neutralino Masses at ILC

The chargino/neutralino pair production is one of the benchmarking processes of ILC. These processes are interesting not only because it allows high precision measurement of chargino and neutralino masses, but also for the reason that the separation of W and Z bosons through their hadronic decay products requires excellent jet resolution being a good benchmark of the detector performance. The analysis based on the SiD detector concept with four jets and missing energy final state will be presented. The uncertainty of chargino and neutralino cross sections can be determined with precision of 0.9% and 4.2% respectively. The mass uncertainties are obtained with a template fitting method achieving precision of better than 1 GeV.

physics.ins-det↗

arXiv2009

The LCFIVertex package: vertexing, flavour tagging and vertex charge reconstruction with an ILC vertex detector

The precision measurements envisaged at the International Linear Collider (ILC) depend on excellent instrumentation and reconstruction software. The correct identification of heavy flavour jets, placing unprecedented requirements on the quality of the vertex detector, will be central for the ILC programme. This paper describes the LCFIVertex software, which provides tools for vertex finding and for identification of the flavour and charge of the leading hadron in heavy flavour jets. These tools are essential for the ongoing optimisation of the vertex detector design for linear colliders such as the ILC. The paper describes the algorithms implemented in the LCFIVertex package, as well as the scope of the code and its performance for a typical vertex detector design.

physics.ins-det↗

arXiv2005

Ultra-High Yield Growth of Vertical Single-Walled Carbon Nanotubes: Hidden Roles of Hydrogen and Oxygen

An oxygen assisted hydrocarbon chemical vapor deposition (CVD) method is developed to afford large-scale highly reproducible ultra high-yield growth of vertical single-walled carbon nanotubes (SWNT). It is revealed that reactive hydrogen (H)-species, inevitable in hydrocarbon-based growth, are damaging to the formation of sp2-like SWNTs. The addition of oxygen scavenges H-species and provides a powerful control over the C/H ratio to favor SWNT growth. The revelation of the roles played by hydrogen and oxygen leads to a unified and universal optimum growth condition for SWNTs. Further, a versatile method is developed to form vertical SWNT films on any substrate, lifting a major substrate-type limitation for aligned SWNTs.

cond-mat.mtrl-sci↗

arXiv2004

On the Origin of Preferential Growth of Semiconducting Single-Walled Carbon Nanotubes

A correlation is observed between the diameter (d) distribution of single walled carbon nanotubes and the percentages of metallic and semiconducting tubes in materials synthesized at low temperature (600 C) by plasma-assisted chemical vapor deposition. Small diameter nanotubes (average d~1.1 nm) show semiconducting-tube percentage much higher than expected for random chirality distribution. Density functional theory calculations reveal discernable differences in the cohesive energies and heat of formation energies for similar-diameter metallic, quasi-metallic and semiconducting nanotubes. Semiconducting nanotubes exhibit the lowest energies and the stabilization effect scales with ~1/d2. This is a likely thermodynamic factor in preferential growth of small diameter semiconducting nanotubes.

cond-mat.mes-hall↗