arXiv subjects

Weizhi Zhong

Publications and source records attributed to Weizhi Zhong.

13 recordsLinked to original sources

arXiv2026

SpatialDiff: 3D-Aware Object Movement via Implicit Spatial Modeling

Recent advances in image editing allow impressive manipulation of objects, existing methods still struggle to handle spatial movement in complex scenes, such as objects span different depth layers or are partially occluded. Most image editing methods focus solely on prior information from 2D datasets, emphasizing planar features while lacking support for spatial structures. Even approaches that incorporate explicit positional information fail to capture true 3D spatial relationships, thus limiting accurate object movement in complex scenes. In this paper, we present SpatialDiff, a method that effectively captures 3D spatial structures, enabling precise and consistent object movements in complex scenes. Our core innovations are twofold: (1) Implicit 3D Spatial Modeling, which introduces 3D prior knowledge and enables the model to internally build a comprehensive understanding of the three-dimensional spatial structure; and (2) Global Spatial Supervision, which constrains the latent spatial features to enable the model to perceive changes in object spatial positions caused by editing operations. Experimental results demonstrate that our method significantly improves the accuracy and fidelity of spatial movement in complex scenes.

cs.CV↗

arXiv2026

SceneExpander: Text-Guided 3D Scene Expansion via Free-Form View Insertion

World building with 3D scene representations is increasingly important for content creation, simulation, and interactive experiences, yet real workflows are inherently iterative: creators repeatedly extend existing scenes under user control. Motivated by this gap, we study text-guided 3D scene expansion via free-form view insertion. Starting from a real scene captured by multi-view images, a user specifies a text expansion intent, which a generative model materializes as an inserted view extending the scene coverage. Unlike simple object editing or style transfer within a fixed scene, the inserted view may be 3D-misaligned with the original reconstruction, introducing geometric shifts, hallucinated content, or view-dependent artifacts that disrupt global multi-view consistency. To address this challenge, we propose SceneExpander, which applies test-time adaptation to a parametric feed-forward 3D reconstruction model with two complementary distillation signals: anchor distillation stabilizes the captured scene using geometric cues from the captured views, while inserted-view self-distillation retains insertion-supported predictions to accommodate the misaligned view. Experiments on ETH scenes and online data demonstrate improved expansion behavior and reconstruction quality under misalignment.

cs.CV↗

arXiv2026

ContextDrag: Precise Drag-Based Image Editing via Context-Preserving Token Injection and Position-Aligned Attention

Drag-based image editing enables intuitive visual manipulation through point-based drag operations. Existing methods mainly rely on diffusion inversion or pixel-space warping with inpainting. However, inversion inherently introduces approximation errors that degrade texture fidelity, whereas rigid pixel-space operations discard semantic context and produce unnatural deformations. To address these issues, we introduce ContextDrag, to our knowledge the first framework that brings drag-based manipulation into the in-context image editing paradigm. By leveraging the in-context capabilities of editing models (e.g., FLUX-Kontext), ContextDrag enables precise drag editing without inversion or fine-tuning. Specifically, we first propose Context-preserving Token Injection (CTI), which injects VAE-encoded reference features into attention layers at spatially aligned target positions, guided by latent-space correspondences estimated directly from user-specified control points. By operating on clean, directly encoded features rather than noisy inversion outputs, CTI preserves rich texture details and enables precise drag control. Second, we propose Position-Aligned Attention (PAA) to eliminate interference caused by spatial displacement of reference features. PAA re-encodes positional embeddings of displaced reference tokens to match their target locations, and masks overlapping regions between source and destination to prevent conflicting features from degrading visual consistency. Experiments on DragBench-SR and DragBench-DR demonstrate that ContextDrag achieves SOTA editing accuracy and overall quality, and comprehensive ablations validate the effectiveness of each proposed component. Code will be publicly available.

cs.CV↗

arXiv2026

Mod-Adapter: Tuning-Free and Versatile Multi-concept Personalization via Modulation Adapter

Personalized text-to-image generation aims to synthesize images of user-provided concepts in diverse contexts. Despite recent progress in multi-concept personalization, most are limited to object concepts and struggle to customize abstract concepts (e.g., pose, lighting). Some methods have begun exploring multi-concept personalization supporting abstract concepts, but they require test-time fine-tuning for each new concept, which is time-consuming and prone to overfitting on limited training images. In this work, we propose a novel tuning-free method for multi-concept personalization that can effectively customize both object and abstract concepts without test-time fine-tuning. Our method builds upon the modulation mechanism in pre-trained Diffusion Transformers (DiTs) model, leveraging the localized and semantically meaningful properties of the modulation space. Specifically, we propose a novel module, Mod-Adapter, to predict concept-specific modulation direction for the modulation process of concept-related text tokens. It introduces vision-language cross-attention for extracting concept visual features, and Mixture-of-Experts (MoE) layers that adaptively map the concept features into the modulation space. Furthermore, to mitigate the training difficulty caused by the large gap between the concept image space and the modulation space, we introduce a VLM-guided pre-training strategy that leverages the strong image understanding capabilities of vision-language models to provide semantic supervision signals. For a comprehensive comparison, we extend a standard benchmark by incorporating abstract concepts. Our method achieves state-of-the-art performance in multi-concept personalization, supported by quantitative, qualitative, and human evaluations.

cs.CV↗

arXiv2025

Style-Preserving Lip Sync via Audio-Aware Style Reference

Audio-driven lip sync has recently drawn significant attention due to its widespread application in the multimedia domain. Individuals exhibit distinct lip shapes when speaking the same utterance, attributed to the unique speaking styles of individuals, posing a notable challenge for audio-driven lip sync. Earlier methods for such task often bypassed the modeling of personalized speaking styles, resulting in sub-optimal lip sync conforming to the general styles. Recent lip sync techniques attempt to guide the lip sync for arbitrary audio by aggregating information from a style reference video, yet they can not preserve the speaking styles well due to their inaccuracy in style aggregation. This work proposes an innovative audio-aware style reference scheme that effectively leverages the relationships between input audio and reference audio from style reference video to address the style-preserving audio-driven lip sync. Specifically, we first develop an advanced Transformer-based model adept at predicting lip motion corresponding to the input audio, augmented by the style information aggregated through cross-attention layers from style reference video. Afterwards, to better render the lip motion into realistic talking face video, we devise a conditional latent diffusion model, integrating lip motion through modulated convolutional layers and fusing reference facial images via spatial cross-attention layers. Extensive experiments validate the efficacy of the proposed approach in achieving precise lip sync, preserving speaking styles, and generating high-fidelity, realistic talking face videos.

cs.CV↗

arXiv2025

UAV-Aided Progressive Interference Source Localization Based on Improved Trust Region Optimization

Trust region optimization-based received signal strength indicator (RSSI) interference source localization methods have been widely used in low-altitude research. However, these methods often converge to local optima in complex environments, degrading the positioning performance. This paper presents a novel unmanned aerial vehicle (UAV)-aided progressive interference source localization method based on improved trust region optimization. By combining the Levenberg-Marquardt (LM) algorithm with particle swarm optimization (PSO), our proposed method can effectively enhance the success rate of localization. We also propose a confidence quantification approach based on the UAV-to-ground channel model. This approach considers the surrounding environmental information of the sampling points and dynamically adjusts the weight of the sampling data during the data fusion. As a result, the overall positioning accuracy can be significantly improved. Experimental results demonstrate the proposed method can achieve high-precision interference source localization in noisy and interference-prone environments.

eess.SP↗

arXiv2024

High-fidelity and Lip-synced Talking Face Synthesis via Landmark-based Diffusion Model

Audio-driven talking face video generation has attracted increasing attention due to its huge industrial potential. Some previous methods focus on learning a direct mapping from audio to visual content. Despite progress, they often struggle with the ambiguity of the mapping process, leading to flawed results. An alternative strategy involves facial structural representations (e.g., facial landmarks) as intermediaries. This multi-stage approach better preserves the appearance details but suffers from error accumulation due to the independent optimization of different stages. Moreover, most previous methods rely on generative adversarial networks, prone to training instability and mode collapse. To address these challenges, our study proposes a novel landmark-based diffusion model for talking face generation, which leverages facial landmarks as intermediate representations while enabling end-to-end optimization. Specifically, we first establish the less ambiguous mapping from audio to landmark motion of lip and jaw. Then, we introduce an innovative conditioning module called TalkFormer to align the synthesized motion with the motion represented by landmarks via differentiable cross-attention, which enables end-to-end optimization for improved lip synchronization. Besides, TalkFormer employs implicit feature warping to align the reference image features with the target motion for preserving more appearance details. Extensive experiments demonstrate that our approach can synthesize high-fidelity and lip-synced talking face videos, preserving more subject appearance details from the reference image.

cs.CV↗

arXiv2023

Identity-Preserving Talking Face Generation with Landmark and Appearance Priors

Generating talking face videos from audio attracts lots of research interest. A few person-specific methods can generate vivid videos but require the target speaker's videos for training or fine-tuning. Existing person-generic methods have difficulty in generating realistic and lip-synced videos while preserving identity information. To tackle this problem, we propose a two-stage framework consisting of audio-to-landmark generation and landmark-to-video rendering procedures. First, we devise a novel Transformer-based landmark generator to infer lip and jaw landmarks from the audio. Prior landmark characteristics of the speaker's face are employed to make the generated landmarks coincide with the facial outline of the speaker. Then, a video rendering model is built to translate the generated landmarks into face images. During this stage, prior appearance information is extracted from the lower-half occluded target face and static reference images, which helps generate realistic and identity-preserving visual content. For effectively exploring the prior information of static reference images, we align static reference images with the target face's pose and expression based on motion fields. Moreover, auditory features are reused to guarantee that the generated face images are well synchronized with the audio. Extensive experiments demonstrate that our method can produce more realistic, lip-synced, and identity-preserving videos than existing person-generic talking face generation methods.

cs.CV↗

arXiv2023

Sparse Bayesian Learning-Based 3D Spectrum Environment Map Construction-Sampling Optimization, Scenario-Dependent Dictionary Construction and Sparse Recovery

The spectrum environment map (SEM), which can visualize the information of invisible electromagnetic spectrum, is vital for monitoring, management, and security of spectrum resources in cognitive radio (CR) networks. In view of a limited number of spectrum sensors and constrained sampling time, this paper presents a new three-dimensional (3D) SEM construction scheme based on sparse Bayesian learning (SBL). Firstly, we construct a scenario-dependent channel dictionary matrix by considering the propagation characteristic of the interested scenario. To improve sampling efficiency, a maximum mutual information (MMI)-based optimization algorithm is developed for the layout of sampling sensors. Then, a maximum and minimum distance (MMD) clustering-based SBL algorithm is proposed to recover the spectrum data at the unsampled positions and construct the whole 3D SEM. We finally use the simulation data of the campus scenario to construct the 3D SEMs and compare the proposed method with the state-of-the-art. The recovery performance and the impact of different sparsity on the constructed SEMs are also analyzed. Numerical results show that the proposed scheme can reduce the required spectrum sensor number and has higher accuracy under the low sampling rate.

eess.SP↗

arXiv2022

A Realistic 3D Non-Stationary Channel Model for UAV-to-Vehicle Communications Incorporating Fuselage Posture

Considering the unmanned aerial vehicle (UAV) three-dimensional (3D) posture, a novel 3D non-stationary geometry-based stochastic model (GBSM) is proposed for multiple-input multiple-output (MIMO) UAV-to-vehicle (U2V) channels. It consists of a line-of-sight (LoS) and non-line-of-sight (NLoS) components. The factor of fuselage posture is considered by introducing a time-variant 3D posture matrix. Some important statistical properties, i.e. the temporal autocorrelation function (ACF) and spatial cross correlation function (CCF), are derived and investigated. Simulation results show that the fuselage posture has significant impact on the U2V channel characteristic and aggravate the non-stationarity. The agreements between analytical, simulated, and measured results verify the correctness of proposed model and derivations. Moreover, it is demonstrated that the proposed model is also compatible to the existing GBSM without considering fuselage posture.

eess.SP↗

arXiv2021

Geometry-Based Stochastic Line-of-Sight Probability Model for A2G Channels under Urban Scenarios

Line-of-sight (LoS) path is essential for the reliability of air-to-ground (A2G) communications, but the existence of LoS path is difficult to predict due to random obstacles on the ground. Based on the statistical geographic information and Fresnel clearance zone, a general stochastic LoS probability model for three-dimensional (3D) A2G channels under urban scenarios is developed. By considering the factors, i.e., building height distribution, building width, building space, carrier frequency, and transceiver's heights, the proposed model is suitable for different frequencies and altitudes. Moreover, in order to get a closed-form expression and reduce the computational complexity, an approximate parametric model is also built with the machine-learning (ML) method to estimate model parameters. The simulation results show that the proposed model has good consistency with existing models at the low altitude. When the altitude increases, it has better performance by comparing with that of the ray-tracing Monte-Carlo simulation data. The analytical results of proposed model are helpful for the channel modeling and performance analysis such as cell coverage, outage probability, and bit error rate in A2G communications.

eess.SP↗

arXiv2021

Map-based Channel Modeling and Generation for U2V mmWave Communication

Unmanned aerial vehicle (UAV) aided millimeter wave (mmWave) technologies have a promising prospect in the future communication networks. By considering the factors of three-dimensional (3D) scattering space, 3D trajectory, and 3D antenna array, a non-stationary channel model for UAV-to-vehicle (U2V) mmWave communications is proposed. The computation and generation methods of channel parameters including interpath and intra-path are analyzed in detail. The inter-path parameters are calculated in a deterministic way, while the parameters of intra-path rays are generated in a stochastic way. The statistical properties are obtained by using a Gaussian mixture model (GMM) on the massive ray tracing (RT) data. Then, a modified method of equal areas (MMEA) is developed to generate the random intra-path variables. Meanwhile, to reduce the complexity of RT method, the 3D propagation space is reconstructed based on the user-defined digital map. The simulated and analyzed results show that the proposed model and generation method can reproduce non-stationary U2V channels in accord with U2V scenarios. The generated statistical properties are consistent with the theoretical and measured ones as well.

eess.SP↗

arXiv2020

Effects of Digital Map on the RT-based Channel Model for UAV mmWave Communications

Based on the geometry and ray tracing (RT) theory, a millimeter wave (mmWave) channel model and parameter computation method for unmanned aerial vehicle (UAV) assisted air-to-ground (A2G) communications are proposed in this paper. In order to speed up the parameter calculation, a reconstruction process of scene database on the original digital map is developed. Moreover, the effects of reconstruction accuracy on the channel parameter and characteristic are analyzed by extensive simulations at 28 GHz under the campus scene. The simulation and analysis results show that the simplified database can save up to 50% time consumption. However, the difference of statistical properties is slight in the campus scenario.

eess.SP↗