Proposes Frequency Dynamic Convolution (FDConv) that learns diverse frequency responses with minimal parameter increase, boosting dense image prediction tasks.

Linwei Chen, Lin Gu, Liang Li et al.

PhysTwin combines spring-mass physics with generative shape models, reconstructing deformable objects from sparse videos with high fidelity.

Hanxiao Jiang, Hao-Yu Hsu, Kaifeng Zhang et al.

Proposes a quantitative evaluation to measure semantic residuals in 3D Gaussian Splatting, validated for privacy-preserving mapping scenarios.

Simona Kocour, Assia Benbihi, Aikaterini Adam et al.

PrediCIR uses a world model to predict missing target features, boosting zero-shot image retrieval by up to 4.45%.

Yuanmin Tang, Jing Yu, Keke Gai et al.

VLU-Net uses a vision-language model for unified multi-degradation image restoration, improving 3.74 dB on the SOTS dehazing dataset.

Haijin Zeng, Xiangming Wang, Yongyong Chen et al.

DCEdit employs dual-level control and precise semantic localization, achieving superior image editing accuracy with no extra training, outperforming existing diffusion-based methods.

Yihan Hu, Jianing Peng, Yiheng Lin et al.

MagicMotion employs dense-to-sparse trajectory guidance within a diffusion transformer framework, outperforming SOTA with a new dataset and benchmark, achieving precise multi-object control.

Quanhao Li, Zhen Xing, Rui Wang et al.

Proposes Gaussian-HOI optimizer for 3D human-object interaction reconstruction; introduces Open3DHOI dataset with 133 object categories and 120 interactions.

Boran Wen, Dingbang Huang, Zichen Zhang et al.

MotionStreamer integrates continuous causal latent space with autoregressive diffusion, enabling real-time streaming human motion synthesis.

Lixing Xiao, Shunlin Lu, Huaijin Pi et al.

EfficientLLaVA employs structural risk minimization for auto-pruning large vision-language models, achieving 83.05% accuracy with only 64 samples and 1.8× speedup.

Yinan Liang, Ziwei Wang, Xiuwei Xu et al.

Proposes TRACE, a semantic-aware transformation method, achieving 30% AUROC improvement for test-time backdoor detection in object detection.

Hangtao Zhang, Yichen Wang, Shihui Yan et al.

Proposes a knowledge-guided LVLM framework achieving 99.53% AUC for deepfake detection, enhancing generalization and explainability.

Peipeng Yu, Jianwei Fei, Hui Gao et al.

DualToken employs dual visual vocabularies to unify understanding and generation, achieving 82.0% zero-shot accuracy on ImageNet with 0.25 rFID.

Wei Song, Yuran Wang, Zijia Song et al.

Concat-ID leverages VAE and 3D self-attention for universal identity-preserving video synthesis across multiple scenarios.

Yong Zhong, Zhuoyi Yang, Jiayan Teng et al.

Introduces multi-output conformal prediction (M-R2CCP, M-R2C2R) for reliable uncertainty quantification in 2D/3D anatomical landmark localization.

Jef Jonkers, Frank Coopman, Luc Duchateau et al.

FlexVLN combines an LLM planner with supervised execution, achieving strong zero-shot transfer across REVERIE, SOON, and CVDN-target.

Siqi Zhang, Yanyuan Qiao, Qunbo Wang et al.

CAVE combines 3D neural object volumes with sparse concept learning, achieving robust out-of-distribution performance and faithful interpretability, with 77.4% OOD accuracy.

Nhi Pham, Artur Jesslen, Bernt Schiele et al.

Proposes On-the-Fly GS, a progressive framework enabling near real-time 3DGS optimization, reducing per-image optimization to seconds.

Yiwei Xu, Yifei Yu, Wentian Gan et al.

Reflect-DiT employs in-context reflection to improve text-to-image diffusion, achieving +0.19 score improvement with only 20 samples on GenEval.

Shufan Li, Konstantinos Kallidromitis, Akash Gokul et al.

Multi-task SceneScript model with infilling enhances 3D scene layout local correction, achieving 98.6% F1 in local tasks.

Christopher Xie, Armen Avetisyan, Henry Howard-Jenkins et al.