5 papers
World Narrative Model for Highly Controllable Video Generation: A Paradigm Shift from Pixel Sampling to Physical World Orchestration
Ye Chen, Xuanhong Chen, Yupeng Zhu +23
The fundamental obstacle to industrial grade video generation is the lack of controllability: existing models treat video as a pixel distribution sampling problem, bypassing the ex…
PASTE: Physics-Aware Scattering Topology Embedding Framework for SAR Object Detection
Jiacheng Chen, Yuxuan Xiong, Haipeng Wang
Current deep learning-based object detection for Synthetic Aperture Radar (SAR) imagery mainly adopts optical image methods, treating targets as texture patches while ignoring inhe…
CEI-3D: Collaborative Explicit-Implicit 3D Reconstruction for Realistic and Fine-Grained Object Editing
Yue Shi, Rui Shi, Yuxuan Xiong +2
Existing 3D editing methods often produce unrealistic and unrefined results due to the deeply integrated nature of their reconstruction networks. To address the challenge, this pap…
HPR3D: Hierarchical Proxy Representation for High-Fidelity 3D Reconstruction and Controllable Editing
Tielong Wang, Yuxuan Xiong, Jinfan Liu +4
Current 3D representations like meshes, voxels, point clouds, and NeRF-based neural implicit fields exhibit significant limitations: they are often task-specific, lacking universal…
DualNeRF: Text-Driven 3D Scene Editing via Dual-Field Representation
Yuxuan Xiong, Yue Shi, Yishun Dou +1
Recently, denoising diffusion models have achieved promising results in 2D image generation and editing. Instruct-NeRF2NeRF (IN2N) introduces the success of diffusion into 3D scene…