arXiv subjects

Xincheng Lin

Publications and source records attributed to Xincheng Lin.

14 recordsLinked to original sources

arXiv2026

MotionMAR: Multi-scale Auto-Regressive Human Motion Reconstruction from Sparse Observations

Human motion follows a temporal hierarchical structure, transitioning from low-frequency global trajectories to high-frequency details. Inspired by the success of multi-level autoregressive models in computer vision, we propose MotionMAR, a coarse-to-fine framework for motion reconstruction from sparse observations. It first estimates the global trajectory of human motion and then gradually refines the temporal details. This architecture consists of four integrated components. The Temporal Multi-scale Tokenization (TMT) VQ-VAE encodes the data at multiple temporal resolutions, separating semantic motion from minor jitters. The Motion Autoregressive Network (MAN) operates in this latent space, predicting motion across scales. It first establishes the global structure through coarse indices and then generates finer indices to recover specific details. Meanwhile, the Scale-Aware Control (SAC) module integrates sparse tracking data to ensure the generated output aligns with actual observations. The Motion Refinement Network (MRN) subsequently smooths consecutive poses and eliminates quantization artifacts. Experiments show that MotionMAR achieves state-of-the-art accuracy on the AMASS dataset, providing a reliable and structure-aware approach for motion reconstruction. The source code is publicly available at http://www.lidarhumanmotion.net/motionmar/.

cs.CV↗

arXiv2026

Three- and four-boson systems expanded around the unitarity limit: Application to $^4$He

The three- and four-boson systems with a large scattering length and a short effective range in the two-body sector are studied in the framework of Short-Range Effective Field Theory. The starting point (leading order) of the EFT is taken to be the universal unitarity limit, where the two-body sector is parameter-free and only one three-body parameter enters. In this limit, physical systems manifests discrete scale invariance. Deviations from universality arising from finite scattering-length and effective-range corrections, as well as a four-body force required by renormalization, are included perturbatively at next-to-leading order. The three-body ground state and its associated four-body ground and first-excited states are studied using the Faddeev-Yakubovsky formalism and a complementary diagrammatic approach. By employing techniques to remove contributions from deep trimers in tetramer calculations, we extend our analysis to larger cutoffs than previously accessible within the FY approach. Our results for binding energies and radii of $^4$He three- and four-atom systems converge well to results obtained with sophisticated phenomenological potentials. These successes suggest that the physics of $^4$He atomic clusters is governed by only small deviations from discrete scale invariance.

cond-mat.quant-gas↗

arXiv2026

SceneParser: Hierarchical Scene Parsing for Visual Semantics Understanding

General scene perception has progressed from object recognition toward open-vocabulary grounding, part localization, and affordance prediction. Yet these capabilities are often realized as isolated predictions that localize objects, parts, or interaction points without capturing the structured dependencies needed for interaction-oriented scene understanding. To address this gap, we introduce Hierarchical Scene Parsing, an interaction-oriented parsing task that represents physical scenes as explicit scene -> object -> part -> affordance hierarchies with cross-level bindings. We instantiate this task with SceneParser, a VLM-based parser trained for unified hierarchical generation with structural-completion pseudo labels and curriculum learning. To support training and evaluation, we construct SceneParser-Bench, a large-scale benchmark built with a scalable hierarchical data engine, containing 110K training images, a 5K validation split, 777K objects, 1.14M parts, 1.74M affordance annotations, and 1.74M valid object-part-affordance chain instances. We further introduce Level-1 to Level-3 conditional metrics and ParseRate to evaluate localization, cross-level binding, and hierarchical completeness. Experiments show that existing MLLMs and perception-stitching pipelines struggle with hierarchical parsing on our SceneParser-Bench, while SceneParser achieves stronger structure-aware performance. Besides, ablations, evaluations on COCO and AGD20K, and a downstream planning probe demonstrate that our SceneParser is compatible with conventional tasks and provides an actionable representation for visual understanding.

cs.CV↗

arXiv2026

Coulomb Effects and Wigner-SU(4) Symmetry in He-3 Charge and Magnetic Properties

This work studies the non-perturbative Coulomb corrections to the He-3 binding energy, magnetic moment, and charge and magnetic radii in leading-order (LO) Pionless Effective Field Theory (Pionless EFT). The splitting between He-3 and H-3 binding energy is found to be 0.85(3) MeV. The Coulomb corrections to the He-3 point charge radius and full magnetic radius are found to be 0.043(2) fm and 0.036(2) fm, respectively. These corrections are 4% of the LO predictions without Coulomb and should be taken into account at next-to-next-to-leading order or beyond in Pionless EFT to achieve the desired EFT accuracy. The Coulomb correction to the He-3 magnetic moment is found to be -0.0041(1)$\mu_N$, only 0.2% of the LO prediction without Coulomb. The impact of Wigner-SU(4) symmetry in the presence of the non-perturbative Coulomb interaction is also discussed and used to help explain the hierarchy of Coulomb effects in He-3 observables.

nucl-th↗

arXiv2026

FlashCap: Millisecond-Accurate Human Motion Capture via Flashing LEDs and Event-Based Vision

Precise motion timing (PMT) is crucial for swift motion analysis. A millisecond difference may determine victory or defeat in sports competitions. Despite substantial progress in human pose estimation (HPE), PMT remains largely overlooked by the HPE community due to the limited availability of high-temporal-resolution labeled datasets. Today, PMT is achieved using high-speed RGB cameras in specialized scenarios such as the Olympic Games; however, their high costs, light sensitivity, bandwidth, and computational complexity limit their feasibility for daily use. We developed FlashCap, the first flashing LED-based MoCap system for PMT. With FlashCap, we collect a millisecond-resolution human motion dataset, FlashMotion, comprising the event, RGB, LiDAR, and IMU modalities, and demonstrate its high quality through rigorous validation. To evaluate the merits of FlashMotion, we perform two tasks: precise motion timing and high-temporal-resolution HPE. For these tasks, we propose ResPose, a simple yet effective baseline that learns residual poses based on events and RGBs. Experimental results show that ResPose reduces pose estimation errors by ~40% and achieves millisecond-level timing accuracy, enabling new research opportunities. The dataset and code will be shared with the community.

cs.CV↗

arXiv2026

Towards Motion Turing Test: Evaluating Human-Likeness in Humanoid Robots

Humanoid robots have achieved significant progress in motion generation and control, exhibiting movements that appear increasingly natural and human-like. Inspired by the Turing Test, we propose the Motion Turing Test, a framework that evaluates whether human observers can discriminate between humanoid robot and human poses using only kinematic information. To facilitate this evaluation, we present the Human-Humanoid Motion (HHMotion) dataset, which consists of 1,000 motion sequences spanning 15 action categories, performed by 11 humanoid models and 10 human subjects. All motion sequences are converted into SMPL-X representations to eliminate the influence of visual appearance. We recruited 30 annotators to rate the human-likeness of each pose on a 0-5 scale, resulting in over 500 hours of annotation. Analysis of the collected data reveals that humanoid motions still exhibit noticeable deviations from human movements, particularly in dynamic actions such as jumping, boxing, and running. Building on HHMotion, we formulate a human-likeness evaluation task that aims to automatically predict human-likeness scores from motion data. Despite recent progress in multimodal large language models, we find that they remain inadequate for assessing motion human-likeness. To address this, we propose a simple baseline model and demonstrate that it outperforms several contemporary LLM-based methods. The dataset, code, and benchmark will be publicly released to support future research in the community.

cs.CV↗

arXiv2025

The Three-Body Limit Cycle: Universal Form for General Regulators

The Efimov effect, a remarkable realization of discrete scale invariance, emerges in the three-body problem with short-range interactions and is understood as a renormalization group (RG) limit cycle within Short-Range Effective Field Theory (SREFT). While the analytic form of the three-body renormalization relation has been established for a sharp cutoff regulator, its universality for other regulators remains underexplored. In this work, we derive the universal functional form of the three-body renormalization relation for general separable regulators through a detailed analysis of the Skorniakov-Ter-Martirosian and Faddeev equations. We find that the relation follows from a real M\"{o}bius transformation characterized by three parameters. This universality is verified numerically for various regulators. Although the functional form remains the same, the parameters characterizing the limit cycle exhibit regulator dependence. These findings broaden the class of RG limit cycles in SREFT and offer a more complete understanding of three-body renormalization.

cond-mat.quant-gas↗

arXiv2025

ClimbingCap: Multi-Modal Dataset and Method for Rock Climbing in World Coordinate

Human Motion Recovery (HMR) research mainly focuses on ground-based motions such as running. The study on capturing climbing motion, an off-ground motion, is sparse. This is partly due to the limited availability of climbing motion datasets, especially large-scale and challenging 3D labeled datasets. To address the insufficiency of climbing motion datasets, we collect AscendMotion, a large-scale well-annotated, and challenging climbing motion dataset. It consists of 412k RGB, LiDAR frames, and IMU measurements, including the challenging climbing motions of 22 skilled climbing coaches across 12 different rock walls. Capturing the climbing motions is challenging as it requires precise recovery of not only the complex pose but also the global position of climbers. Although multiple global HMR methods have been proposed, they cannot faithfully capture climbing motions. To address the limitations of HMR methods for climbing, we propose ClimbingCap, a motion recovery method that reconstructs continuous 3D human climbing motion in a global coordinate system. One key insight is to use the RGB and LiDAR modalities to separately reconstruct motions in camera coordinates and global coordinates and to optimize them jointly. We demonstrate the quality of the AscendMotion dataset and present promising results from ClimbingCap. The AscendMotion dataset and source code release publicly at \href{this link}{http://www.lidarhumanmotion.net/climbingcap/}

cs.CV↗

arXiv2024

Two-Body Triton Photodisintegration and Wigner-SU(4) Symmetry

We calculate the two-body triton photodisintegration cross section as a function of photon energy to next-to-next-to leading order (NNLO) in pionless effective field theory (EFT($\pi \!\!/$)) and show good agreement with experiment. In addition we calculate the polarization asymmetry $R_C=-0.441(15)$ in cold neutron-deuteron capture to NNLO in EFT($\pi \!\!/$), in agreement with the experimental value of $R_C=-0.42\pm 0.03$ [M. W. Konijnenberg et al. in Phys. Lett. B 205, 215 (1988)]. We also assess the dependence of $R_C$ on different fits of the two-nucleon magnetic currents. Finally, we consider the impact of Wigner-SU(4) symmetry and demonstrate that starting from the Wigner-SU(4) symmetric limit and including perturbative corrections to the breaking of Wigner-SU(4) symmetry does a good job of describing two-body triton photodisintegration.

nucl-th↗

arXiv2024

RELI11D: A Comprehensive Multimodal Human Motion Dataset and Method

Comprehensive capturing of human motions requires both accurate captures of complex poses and precise localization of the human within scenes. Most of the HPE datasets and methods primarily rely on RGB, LiDAR, or IMU data. However, solely using these modalities or a combination of them may not be adequate for HPE, particularly for complex and fast movements. For holistic human motion understanding, we present RELI11D, a high-quality multimodal human motion dataset involves LiDAR, IMU system, RGB camera, and Event camera. It records the motions of 10 actors performing 5 sports in 7 scenes, including 3.32 hours of synchronized LiDAR point clouds, IMU measurement data, RGB videos and Event steams. Through extensive experiments, we demonstrate that the RELI11D presents considerable challenges and opportunities as it contains many rapid and complex motions that require precise location. To address the challenge of integrating different modalities, we propose LEIR, a multimodal baseline that effectively utilizes LiDAR Point Cloud, Event stream, and RGB through our cross-attention fusion strategy. We show that LEIR exhibits promising results for rapid motions and daily motions and that utilizing the characteristics of multiple modalities can indeed improve HPE performance. Both the dataset and source code will be released publicly to the research community, fostering collaboration and enabling further exploration in this field.

cs.CV↗

arXiv2024

The EDGE-CALIFA Survey: An Extragalactic Database for Galaxy Evolution Studies

The EDGE-CALIFA survey provides spatially resolved optical integral field unit (IFU) and CO spectroscopy for 125 galaxies selected from the CALIFA Data Release 3 sample. The Extragalactic Database for Galaxy Evolution (EDGE) presents the spatially resolved products of the survey as pixel tables that reduce the oversampling in the original images and facilitate comparison of pixels from different images. By joining these pixel tables to lower dimensional tables that provide radial profiles, integrated spectra, or global properties, it is possible to investigate the dependence of local conditions on large-scale properties. The database is freely accessible and has been utilized in several publications. We illustrate the use of this database and highlight the effects of CO upper limits on the inferred slopes of the local scaling relations between stellar mass, star formation rate (SFR), and H$_2$ surface densities. We find that the correlation between H$_2$ and SFR surface density is the tightest among the three relations.

astro-ph.GA↗

arXiv2023

Four-Body Systems at Large Cutoffs in Effective Field Theory

Four-body systems are studied using an effective field theory with two- and three-body contact interactions. A method to systematically address deep trimers (three-body bound states that are more tightly bound than four-body bound/resonant states) in four-body calculations is developed using a diagrammatic approach. Previous calculations were limited by the existence of deep trimers, which this work overcomes. For cold $^4$He atoms, binding energies of 526.1(5)~mK and 128.517(1)~mK are obtained at leading order for the tetramer ground and excited states, respectively, where errors come from the truncation of three-body partial waves. Tetramer binding energies and decay widths are also computed approaching the unitary limit. In the unitary limit, there are two tetramers associated with each trimer of binding energy $B_3^{(0)}$. The binding energy and decay width for the associated tetramer ground state are $E_4^{(0)} = 4.60(1)B_3^{(0)}$ and $\Gamma_4^{(0)}/2 = 0.0160(1)B_3^{(0)}$, respectively, and for the associated tetramer excited state, $E_4^{(1)} = 1.0022(3)B_3^{(0)}$ and $\Gamma_4^{(1)}/2 = 2.57(2)\times 10^{-4}B_3^{(0)}$, respectively. This calculation is a gateway to higher-order and/or more-body calculations in nuclear and atomic systems.

nucl-th↗

arXiv2022

Cold Neutron-Deuteron Capture and Wigner-SU(4) Symmetry

We calculate the cold neutron-deuteron ($nd$) capture cross section, $\sigma_{nd}$, to next-to-next-to leading order (NNLO) using the model-independent approach of pionless effective field theory (EFT($\pi\!\!\!/$)). At leading order we find $\sigma_{nd} = 0.315 \pm 0.217$ mb, while the experimental result is 0.508(15) mb [Jurney, Bendt and Browne in Phys. Rev. C 25, 2810 (1982)] for a laboratory neutron velocity of 2200 m/s. At next-to-leading-order (NLO), we show that $\sigma_{nd}$ is sensitive to the low energy constant (LEC), $L_1^{(0)}$, of the two-nucleon isovector current appearing at NLO. A fit of $L_1^{(0)}$ at NLO to the triton magnetic moment yields a NLO prediction of $\sigma_{nd}=0.393 \pm 0.164$ mb, where the error comes from propagating the error from the $L_1^{(0)}$ fit. At next-to-next-to-leading-order (NNLO), we find that a new three-nucleon magnetic moment counterterm is required for renormalization group invariance of both $\sigma_{nd}$ and the triton magnetic moment. Fitting the NNLO correction to $L_1^{(0)}$ (denoted $L_1^{(1)}$) to cold neutron-proton capture ($\sigma_{np}$) yields a NNLO prediction of $\sigma_{nd}=0.447 \pm 0.130$ mb, where the error comes from propagating the error from the $L_1^{(1)}$ fit. We also study different fittings of $L_1^{(0)}$ and $L_1^{(1)}$ to $\sigma_{np}$, $\sigma_{nd}$, and/or the triton magnetic moment. For example, fitting $L_1^{(0)}$ simultaneously to $\sigma_{np}$, $\sigma_{nd}$, and the triton magnetic moment at NLO, and fitting $L_1^{(1)}$ simultaneously to $\sigma_{np}$ and $\sigma_{nd}$ at NNLO, yields $\sigma_{nd} = 0.480 \pm 0.114$ mb and $0.511 \pm 0.042$ mb, respectively, where errors are naively estimated from EFT($\pi\!\!\!/$) power counting. In addition, we discuss how Wigner-SU(4) symmetry may alter the naive EFT($\pi\!\!\!/$) expansion of $\sigma_{nd}$.

nucl-th↗

arXiv2022

Large-$N_c$ constraints for elastic dark matter-light nucleus scattering in pionless effective field theory

Recent proposals for the use of light nuclei as dark matter direct detection targets necessitate a strong theoretical understanding of the nuclear physics involved. We perform relevant calculations for dark matter-light nucleus scattering in a combined pionless effective field theory and large-$N_c$ expansion, where $N_c$ is the number of quark colors. We include a general set of one-nucleon currents that have been used in other effective theories, as well as novel two-nucleon contact currents. First, we obtain constraints for the relative sizes of the dark matter couplings to the one- and two-nucleon currents through the large-$N_c$ expansion. Then, we use these constraints to make predictions for the relative sizes of spin-dependent and spin-independent cross sections for dark matter scattering off of a nucleon, a deuteron, a triton, and helium-3.

nucl-th↗