arXiv subjects

Jianxin Huang

Publications and source records attributed to Jianxin Huang.

8 recordsLinked to original sources

arXiv2026

USR-Drive: Unified Driving Scene Representation via Joint Denoising of 3D Gaussians and Boxes

Spatial representation learning for autonomous driving aims to map raw visual signals into structured 3D scene representations, where object-centric bounding boxes and rendering-oriented 3D primitives (\eg, 3D Gaussians) serve as two distinct yet highly complementary levels for scene understanding. Existing methods typically treat dynamic reconstruction and instance-level perception as separate tasks, despite their shared goal of estimating the underlying 3D world state. As a result, dynamic reconstruction is under-constrained while 3D detection lacks geometric grounding. To address this gap, we propose USR-Drive, a unified conditional generative framework that, given only posed multi-view driving videos, jointly recovers dense dynamic geometry and instance-level object layouts within a shared scene representation. Specifically, USR-Drive represents dense Gaussian primitives and sparse 3D bounding boxes as two aligned latent token streams and jointly denoises them with a unified multi-modal diffusion Transformer. Unlike prior paradigms that use boxes as external conditions or predict them with detached modules, USR-Drive treats them as mutually constrained state variables with a Unified Positional Encoding (UPE) that aligns heterogeneous tokens within a shared metric spatiotemporal coordinate. Via such unified representation and generative framework, the two modalities reinforce each other: geometry supplies dense metric evidence for box prediction, while boxes provide instance-level structural priors that help preserve spatial consistency and reduce ambiguity in sequential 3D geometric representation. Our approach successfully delivers state-of-the-art results for both dynamic reconstruction and 3D detection on the nuScenes and VKitti datasets.

cs.CV↗

arXiv2026

Prominent Intrinsic Orbital Hall Effect in the 135 Kagome Metal Family and Orbital Responses in the Loop-Current Phase

The study of electronic orbital degrees of freedom, including the generation and control of orbital currents and orbital angular momentum, has emerged as a vibrant research field. Here, we study the orbital Hall effect (OHE), one of the key mechanisms for orbital current generation, in transition-metal kagome-lattice metals. We propose a large positive OHE in CsTi$_3$Bi$_5$ and negative OHEs in CsV$_3$Sb$_5$ and CsCr$_3$Sb$_5$ models. Orbital-sector decomposition shows that the $\vert l^z_d \vert=2$ $d$-orbital channel gives a positive contribution, whereas the $\vert l^z \vert=1$ $d$- and $p$-orbital channels can give negative contributions. Control calculations suggest a persistent orbital-sector sign tendency near the actual filling, while strong $p$-$d$ hybridization modulates the quantitative balance and total OHE sign. Thus, the compound and filling dependence of the OHE reflects both the sign tendency of each orbital sector and the hybridization-controlled balance among them. Furthermore, we investigate the loop-current phase of CsV$_3$Sb$_5$ and show that it induces finite local atomic orbital angular momentum. We also show that loop-current-related symmetry lowering allows finite symmetric components of the orbital conductivity tensor. This study provides a basis for exploring orbital currents and local orbital-angular-momentum responses in strongly correlated kagome metals.

cond-mat.str-el↗

arXiv2026

Gaussian-LIC2: LiDAR-Inertial-Camera Gaussian Splatting SLAM

This paper presents the first photo-realistic LiDAR-Inertial-Camera Gaussian Splatting SLAM system that simultaneously addresses visual quality, geometric accuracy, and real-time performance. The proposed method performs robust and accurate pose estimation within a continuous-time trajectory optimization framework, while incrementally reconstructing a 3D Gaussian map using camera and LiDAR data, all in real time. The resulting map enables high-quality, real-time novel view rendering of both RGB images and depth maps. To effectively address under-reconstruction in regions not covered by the LiDAR, we employ a lightweight zero-shot depth model that synergistically combines RGB appearance cues with sparse LiDAR measurements to generate dense depth maps. The depth completion enables reliable Gaussian initialization in LiDAR-blind areas, significantly improving system applicability for sparse LiDAR sensors. To enhance geometric accuracy, we use sparse but precise LiDAR depths to supervise Gaussian map optimization and accelerate it with carefully designed CUDA-accelerated strategies. Furthermore, we explore how the incrementally reconstructed Gaussian map can improve the robustness of odometry. By tightly incorporating photometric constraints from the Gaussian map into the continuous-time factor graph optimization, we demonstrate improved pose estimation under LiDAR degradation scenarios. We also showcase downstream applications via extending our elaborate system, including video frame interpolation and fast 3D mesh extraction. To support rigorous evaluation, we construct a dedicated LiDAR-Inertial-Camera dataset featuring ground-truth poses, depth maps, and extrapolated trajectories for assessing out-of-sequence novel view synthesis. Both the dataset and code will be made publicly available on project page https://xingxingzuo.github.io/gaussian_lic2.

cs.RO↗

arXiv2025

DepthMatch: Semi-Supervised RGB-D Scene Parsing through Depth-Guided Regularization

RGB-D scene parsing methods effectively capture both semantic and geometric features of the environment, demonstrating great potential under challenging conditions such as extreme weather and low lighting. However, existing RGB-D scene parsing methods predominantly rely on supervised training strategies, which require a large amount of manually annotated pixel-level labels that are both time-consuming and costly. To overcome these limitations, we introduce DepthMatch, a semi-supervised learning framework that is specifically designed for RGB-D scene parsing. To make full use of unlabeled data, we propose complementary patch mix-up augmentation to explore the latent relationships between texture and spatial features in RGB-D image pairs. We also design a lightweight spatial prior injector to replace traditional complex fusion modules, improving the efficiency of heterogeneous feature fusion. Furthermore, we introduce depth-guided boundary loss to enhance the model's boundary prediction capabilities. Experimental results demonstrate that DepthMatch exhibits high applicability in both indoor and outdoor scenes, achieving state-of-the-art results on the NYUv2 dataset and ranking first on the KITTI Semantics benchmark.

cs.CV↗

arXiv2025

Odd-parity quadrupole order and induced nonreciprocal transport in the kagome metal CsTi$_3$Bi$_5$ driven by quantum interference

Kagome metals present a fascinating platform of quantum phases thanks to the interplay between the geometric frustration and strong electron correlation. Here, we propose the emergence of the electric odd-parity bond order (BO) that originates from the intra-unit-cell odd-parity configuration in recently discovered kagome metal CsTi$_3$Bi$_5$. The predicted E1u BO is induced by the beyond-mean-field mechanism, that is, the quantum interference among different sublattice spin fluctuations. Importantly, the accompanied nematic deformation of the Fermi surface is just ~1% while the intensity of the quasiparticle interference signal exhibits drastic nematic anisotropy, consistent with the scanning tunneling microscope measurements in CsTi$_3$Bi$_5$. The present odd-parity BO triggers interesting phenomena, such as the non-linear Hall effect and emergent electromagnetism.

cond-mat.str-el↗

arXiv2024

RoadFormer+: Delivering RGB-X Scene Parsing through Scale-Aware Information Decoupling and Advanced Heterogeneous Feature Fusion

Task-specific data-fusion networks have marked considerable achievements in urban scene parsing. Among these networks, our recently proposed RoadFormer successfully extracts heterogeneous features from RGB images and surface normal maps and fuses these features through attention mechanisms, demonstrating compelling efficacy in RGB-Normal road scene parsing. However, its performance significantly deteriorates when handling other types/sources of data or performing more universal, all-category scene parsing tasks. To overcome these limitations, this study introduces RoadFormer+, an efficient, robust, and adaptable model capable of effectively fusing RGB-X data, where ``X'', represents additional types/modalities of data such as depth, thermal, surface normal, and polarization. Specifically, we propose a novel hybrid feature decoupling encoder to extract heterogeneous features and decouple them into global and local components. These decoupled features are then fused through a dual-branch multi-scale heterogeneous feature fusion block, which employs parallel Transformer attentions and convolutional neural network modules to merge multi-scale features across different scales and receptive fields. The fused features are subsequently fed into a decoder to generate the final semantic predictions. Notably, our proposed RoadFormer+ ranks first on the KITTI Road benchmark and achieves state-of-the-art performance in mean intersection over union on the Cityscapes, MFNet, FMB, and ZJU datasets. Moreover, it reduces the number of learnable parameters by 65\% compared to RoadFormer. Our source code will be publicly available at mias.group/RoadFormerPlus.

cs.CV↗

arXiv2024

Low temperature phase transitions under CDW state in kagome metals AV3Sb5 (A=Cs,Rb,K): Significance of mix-type Fermi surface electron correlations

To understand the multistage phase transitions in V-based kagome metals inside the charge-density-wave (CDW) phase, we focus on the impact of the "mix-type" Fermi surface because it is intact in the CDW state on the "pure-type" Fermi surface. On the mixed-type Fermi surface, moderate spin correlations develop, and we reveal that uniform (q=0) bond order is caused by the paramagnon interference mechanism, which is described by the Aslamazov-Larkin vertex correction. A dominant solution is the E2g-symmetry nematic order, in which the director can be rotated arbitrarily. In addition, we obtain the A1g-symmetry non-nematic order, which leads to the change in the lattice constants without symmetry breaking. The predicted E2g and A1g channel fluctuations at q=0 can be observed by the elastoresistance measurements. These results are useful to understand the multistage phase transitions inside the 2 $\times$ 2 CDW phase. The present theory has a general significance because mix-type band structure universally exists in various kagome lattice models.

cond-mat.str-el↗

arXiv2022

Ctrl-VIO: Continuous-Time Visual-Inertial Odometry for Rolling Shutter Cameras

In this paper, we propose a probabilistic continuous-time visual-inertial odometry (VIO) for rolling shutter cameras. The continuous-time trajectory formulation naturally facilitates the fusion of asynchronized high-frequency IMU data and motion-distorted rolling shutter images. To prevent intractable computation load, the proposed VIO is sliding-window and keyframe-based. We propose to probabilistically marginalize the control points to keep the constant number of keyframes in the sliding window. Furthermore, the line exposure time difference (line delay) of the rolling shutter camera can be online calibrated in our continuous-time VIO. To extensively examine the performance of our continuous-time VIO, experiments are conducted on publicly-available WHU-RSVI, TUM-RSVI, and SenseTime-RSVI rolling shutter datasets. The results demonstrate the proposed continuous-time VIO significantly outperforms the existing state-of-the-art VIO methods. The codebase of this paper will also be open-sourced at \url{https://github.com/APRIL-ZJU/Ctrl-VIO}.

cs.RO↗