arXiv subjects

Liang Lei

Publications and source records attributed to Liang Lei.

5 recordsLinked to original sources

arXiv2026

SimFuse3D: Source-Guided Target Simulation and Confidence-Guided Multi-Stage Localization Reweighting for Cross-Platform 3D Object Detection

Changes in sensor height and viewpoint alter object-level point distributions, making cross-platform LiDAR unsupervised domain adaptation (UDA) difficult. Self-training uses labeled source scans and unlabeled target scans, yet a retained prediction may provide a useful target location while enclosing sparse foreground returns, background clutter, or points inconsistent with the predicted box. We refer to this mismatch as box-point inconsistency. We introduce SimFuse3D, which preserves the target placement and repairs the associated pseudo-object using measured geometry from labeled source scans. Object Memory retrieves a compatible labeled source instance. Target Simulation places its ground-truth box at the target location, aligns its points with the target viewing geometry, and filters the aligned crop to approximate the target observation. Confidence-Guided Multi-Stage Localization Reweighting (CMLR) maps each target pseudo-object confidence score to a bounded weight shared by RPN localization and R-CNN box regression. All components operate only during adaptation, leaving the detector architecture and inference graph unchanged. Across six cross-platform transfers, SimFuse3D exceeds Pi3DET-Net on every reported AP metric and ranks first among the compared adaptation methods on nearly all metrics. On nuScenes-to-KITTI, it ranks first among the compared adaptation methods with both evaluated detectors.

cs.CV↗

arXiv2026

Centimeter-scale fully suspended metal and metal oxide thin films by one-step transfer-free liquid metal capillary forming

Fully suspended thin films can decouple substrate effects and provide additional tuning degrees of freedom compared with their substrate-supported counterparts, making them unique platforms for next-generation thin film devices. Here we report one-step, transfer-free and substrate-free fabrication of centimeter-scale ultrathin fully suspended metal and metal oxide film structures via liquid metal capillary forming. We show that, analogous to soap film formation, the instantaneously developed few-nanometer-thick native surface oxide can laminate various liquid metals into micrometer-thick metallic films. Surprisingly, the surfactant-like metal oxide bilayer can survive dewetting-induced liquid metal drainage, forming suspended two-dimensional films featuring an enormous lateral size-to-thickness ratio on the order of 10^7. We further demonstrate rapid prototyping of metallic minimal-surface thin-walled structures and ultra-sensitive acoustic wave detection with these suspended thin film platforms.

cond-mat.mtrl-sci↗

arXiv2025

Achieve equilibrium outside the contact angle hysteresis

It is common belief that the equilibrium contact angle, corresponding to the minimum system energy state, lies between advancing and receding contact angles. Here, we derive advancing and receding contact angles considering the micro contacting processes on ideal rough 2D surfaces. Equilibrium contact angles obtained via energy minimization can be smaller than the receding contact angle and reach 0 degrees, at which hysteresis diminishes on the super-hydrophilic surface. Gibbs free energy analyses, numerical simulations and physical experiments all confirm these new findings.

cond-mat.soft↗

arXiv2023

Characterization and Generation of 3D Realistic Geological Particles with Metaball Descriptor based on X-Ray Computed Tomography

The morphology of geological particles is crucial in determining its granular characteristics and assembly responses. In this paper, Metaball-function based solutions are proposed for morphological characterization and generation of three-dimensional realistic particles according to the X-ray Computed Tomography (XRCT) images. For characterization, we develop a geometric-based Metaball-Imaging algorithm. This algorithm can capture the main contour of parental particles with a series of non-overlapping spheres and refine surface-texture details through gradient search. Four types of particles, hundreds of samples, are applied for evaluations. The result shows good matches on key morphological indicators(i.e., volume, surface area, sphericity, circularity, corey-shape factor, nominal diameter and surface-equivalent-sphere diameter), confirming its characterization precision. For generation, we propose the Metaball Variational Autoencoder. Assisted by deep neural networks, this method can generate 3D particles in Metaball form, while retaining coessential morphological features with parental particles. Additionally, this method allows for control over the generated shapes through an arithmetic pattern, enabling the generation of particles with specific shapes. Two sets of XRCT images different in sample number and geometric features are chosen as parental data. On each training set, one thousand particles are generated for validations. The generation fidelity is demonstrated through comparisons of morphologies and shape-feature distributions between generated and parental particles. Examples are also provided to demonstrate controllability on the generated shapes. With Metaball-based simulations frameworks previously proposed by the authors, these methods have the potential to provide valuable insights into the properties and behavior of actual geological particles.

eess.IV↗

arXiv2022

Metaball-Imaging Discrete Element Lattice Boltzmann Method for fluid-particle system of complex morphologies with settling case study

Fluid-particle systems are highly sensitive to particle morphologies. While many attempts have been made on shape descriptors and coupling schemes, how to simulate the particle-particle and particle-fluid interactions with a balance between accuracy and efficiency is still a challenge, especially when complex-shaped particles are considered. This study presents a Metaball-Imaging (MI) based Discrete Element Lattice Boltzmann Method (DELBM) for fluid simulations with irregular shaped particles. The major innovation is the MI algorithm to capture the real grain shape for DELBM simulations,where the Metaball function is utilized as the mathematical representation due to its versatile and efficient expressiveness of complex shapes.The contact detection is tackled robustly by gradient calculation of the closest point with a Newton-Raphson based scheme. And the coupling with LBM is accomplished by a classic sharp-interface scheme. As for refiling, a local refiling algorithm based on the bounce back rule is implemented. Validations on three settling experiments of irregular-shaped natural cobblestones indicate the proposed model to be effective and powerful in probing micromechanics of irregular-shaped granular media immersed in fluid systems. The potential of this model on studies of shape-induced physical processes is further investigated with numerical examples on the "drafting, kissing and tumbling" phenomenon of pair particles in various shapes.

cs.CE↗