arXiv subjects

Wenlong Cheng

Publications and source records attributed to Wenlong Cheng.

7 recordsLinked to original sources

arXiv2026

FedOT: Ownership Verification and Leakage Tracing via Watermarks for Federated LDMs

Training Latent Diffusion Models (LDMs) within Federated Learning (FL) has attracted increasing attention due to its ability to combine the powerful generative capacity of LDMs with the privacy-preserving properties of FL. However, FL requires sharing the global model with multiple participants, which risks unauthorized model distribution or resale by malicious clients. While an intuitive approach is to adopt existing VAE-based watermarking techniques for LDMs in FL, this strategy falls short in addressing such threats due to two fundamental challenges: (1) Existing methods support ownership verification but lack the ability to trace model leakage to a specific malicious client; (2) VAE-based watermarks are vulnerable, as they can be removed simply by replacing the decoder with a clean counterpart. In this paper, we propose FedOT, the first framework for ownership verification and leakage tracing in federated LDMs. Specifically, to address the first challenge, we design a chunked watermark, where the first part is for ownership verification, and the second part is used for client identification. Furthermore, to overcome the second challenge and secure the model against VAE replacement attack, we introduce Latent Vector Transformation (LVT), which strengthens the connection between the VAE and U-Net latent spaces by modifying the original latent distribution of the VAE. Consequently, any attempt to replace the VAE for watermark removal leads to significant image quality degradation, making the LDM model unusable. Extensive experiments demonstrate that FedOT achieves superior performance in both ownership verification and traceability. Project page: https://spyzixuan.github.io/FedOT/.

cs.CV↗

arXiv2025

Towards Objective Obstetric Ultrasound Assessment: Contrastive Representation Learning for Fetal Movement Detection

Accurate fetal movement (FM) detection is essential for assessing prenatal health, as abnormal movement patterns can indicate underlying complications such as placental dysfunction or fetal distress. Traditional methods, including maternal perception and cardiotocography (CTG), suffer from subjectivity and limited accuracy. To address these challenges, we propose Contrastive Ultrasound Video Representation Learning (CURL), a novel self-supervised learning framework for FM detection from extended fetal ultrasound video recordings. Our approach leverages a dual-contrastive loss, incorporating both spatial and temporal contrastive learning, to learn robust motion representations. Additionally, we introduce a task-specific sampling strategy, ensuring the effective separation of movement and non-movement segments during self-supervised training, while enabling flexible inference on arbitrarily long ultrasound recordings through a probabilistic fine-tuning approach. Evaluated on an in-house dataset of 92 subjects, each with 30-minute ultrasound sessions, CURL achieves a sensitivity of 78.01% and an AUROC of 81.60%, demonstrating its potential for reliable and objective FM analysis. These results highlight the potential of self-supervised contrastive learning for fetal movement analysis, paving the way for improved prenatal monitoring and clinical decision-making.

cs.CV↗

arXiv2021

MFAGAN: A Compression Framework for Memory-Efficient On-Device Super-Resolution GAN

Generative adversarial networks (GANs) have promoted remarkable advances in single-image super-resolution (SR) by recovering photo-realistic images. However, high memory consumption of GAN-based SR (usually generators) causes performance degradation and more energy consumption, hindering the deployment of GAN-based SR into resource-constricted mobile devices. In this paper, we propose a novel compression framework \textbf{M}ulti-scale \textbf{F}eature \textbf{A}ggregation Net based \textbf{GAN} (MFAGAN) for reducing the memory access cost of the generator. First, to overcome the memory explosion of dense connections, we utilize a memory-efficient multi-scale feature aggregation net as the generator. Second, for faster and more stable training, our method introduces the PatchGAN discriminator. Third, to balance the student discriminator and the compressed generator, we distill both the generator and the discriminator. Finally, we perform a hardware-aware neural architecture search (NAS) to find a specialized SubGenerator for the target mobile phone. Benefiting from these improvements, the proposed MFAGAN achieves up to \textbf{8.3}$\times$ memory saving and \textbf{42.9}$\times$ computation reduction, with only minor visual quality degradation, compared with ESRGAN. Empirical studies also show $\sim$\textbf{70} milliseconds latency on Qualcomm Snapdragon 865 chipset.

cs.AR↗

arXiv2020

Direct imaging of liquid-nanoparticle interface with atom probe tomography

Understanding the structure and chemical composition at the liquid-nanoparticle (NP) interface is crucial for a wide range of physical, chemical and biological processes. In this study, direct imaging of the liquid-NP interface by atom probe tomography (APT) is reported for the first time, which reveals the distributions and the interactions of key atoms and molecules in this critical domain. The APT specimen is prepared by controlled graphene encapsulation of the solution containing nanoparticles on a metal tip, with an end radius in the range of 50 nm to allow field ionization and evaporation. Using Au nanoparticles (AuNPs) in suspension as an example, analysis of the mass spectrum and three-dimensional (3D) chemical maps from APT provides a detailed image of the water-gold interface with near-atomic resolution. At the water-gold interface, the formation of an electrical double layer (EDL) rich in water (H2O) molecules has been observed, which results from the charge from the binding between the trisodium-citrate layer and the AuNP. In the bulk water region, the density of reconstructed H2O has been shown to be consistent, reflecting a highly packed density of H2O molecules after graphene encapsulation. This study is the first demonstration of direct imaging of liquid-NP interface using APT with results providing an atom-by-atom 3D dissection of the liquid-NP interface.

cond-mat.mtrl-sci↗

arXiv2019

An Adaptive Soft Plasmonic Nanosheet Resonator

Current micro nanomechanical system are usually based on rigid crystalline semiconductors that normally have high quality factors but lack adaptive responses to variable frequencies, a capability ubiquitous for communications in the biological world, such as bat and whale calls. Here, we demonstrate a soft mechanical resonator based on a freestanding organic-inorganic hybrid plasmonic superlattice nanosheet, which can respond adaptively to either incident light intensity or wavelength. This is achieved because of strong plasmonic coupling in closely-packed nanocrystals which can efficiently concentrate and convert photons into heat. The heat causes the polymer matrix to expand, leading to a change in the nanomechanical properties of the plasmonic nanosheet. Notably, the adaptive frequency responses are also reversible and the responsive ranges are fine-tunable by adjusting the constituent nanocrystal building blocks. We believe that our plasmonic nanosheets may open a new route to design next-generation intelligent bio-mimicking opto-mechanical resonance systems.

physics.app-ph↗

arXiv2018

Near-Perfect Absorption of Light by Plasmene Sheets

Near-perfect absorbers (NPAs) efficiently absorb visible light with a layered nanostructure that is thinner than the diffusion lengths of photogenerated charge carriers. We overcame existing limitations in fabricating their nanoparticulate surface by depositing \textit{plasmene}, a tightly-packed two-dimensional lattice of metal nanoparticles formed through self-assembly. The plasmene NPAs absorb up to 98\% of incident visible light, with modelling showing the improvement on existing NPAs arises from the structural ordering of the plasmene. We also demonstrate control of NPAs' absorption profile through the use of anisotropic building blocks in plasmene. These property enhancements may broaden the application of NPAs to structural colour, sensing and photocatalysis.

physics.app-ph↗

arXiv2015

Extensional viscosity of copper nanowire suspensions in an aqueous polymer solution

Suspensions of copper nanowires are emerging as new electronic inks for next-generation flexible electronics. Using a novel surface acoustic wave driven extensional flow technique we are able to perform currently lacking analysis of these suspensions and their complex buffer. We observe extensional viscosities from 3 mPa$\cdot$s (1 mPa$\cdot$s shear viscosity) to 37.2 Pa$\cdot$s via changes in the suspension concentration, thus capturing low viscosities that have been historically very challenging to measure. These changes equate to an increase in the relative extensional viscosity of nearly 12,200 times at a volume fraction of just 0.027. We also find that interactions between the wires and the necessary polymer additive affect the rheology strongly. Polymer-induced elasticity shows a reduction as the buffer relaxation time falls from 819 to 59 $μ$s above a critical particle concentration. The results and technique presented here should aid in the future formulation of these promising nanowire suspensions and their efficient application as inks and coatings.

cond-mat.soft↗