5 papers
A Degradation-Tolerance Benchmark for Camera-Only End-to-End Driving
Haohua Que, Handong Yao
Camera-only end-to-end (E2E) driving models are nearing deployment, where the camera stream is degraded by blur, noise, low light, weather, frame loss, and memory faults. How much…
DinoLink: A Token-Centric Representation Compression Framework for Bandwidth-Constrained Collaborative V2X Perception
Tianle Zhu, Haohua Que, Handong Yao +2
High-precision remote perception is often hindered by the severe bandwidth constraints of Vehicle-to-Everything (V2X) networks. We propose \textit{DinoLink}, a token-centric compre…
ACEsplat: Accelerated 3D Gaussian Scene Regression via RGB and Poses Only
Mingkai Liu, Haohua Que, Dikai Fan +7
Per-scene 3D Gaussian Splatting (3DGS) enables high-fidelity rendering, but practical robotic and AR scene capture pipelines often depend on external geometric initialization (e.g.…
CABLE: Cloud-Assisted Bandwidth-efficient LMM-based Encoding for V2X Systems
Haohua Que, Zhipeng Bao, Qianyi Wu +1
Cloud-hosted large multimodal models (LMMs) can provide strong open-vocabulary perception for Vehicle-to-Everything systems, but naively transmitting full-resolution frames from ed…
MACE: Mixture-of-Experts Accelerated Coordinate Encoding for Large-Scale Scene Localization and Rendering
Mingkai Liu, Dikai Fan, Haohua Que +10
Efficient localization and high-quality rendering in large-scale scenes remain a significant challenge due to the computational cost involved. While Scene Coordinate Regression (SC…