Zobrazeno 1 - 10
of 77
pro vyhledávání: '"Xie, Yichen"'
Autor:
Peng, Chensheng, Zhang, Chengwei, Wang, Yixiao, Xu, Chenfeng, Xie, Yichen, Zheng, Wenzhao, Keutzer, Kurt, Tomizuka, Masayoshi, Zhan, Wei
We present DeSiRe-GS, a self-supervised gaussian splatting representation, enabling effective static-dynamic decomposition and high-fidelity surface reconstruction in complex driving scenarios. Our approach employs a two-stage optimization pipeline o
Externí odkaz:
http://arxiv.org/abs/2411.11921
Autor:
Xie, Yichen, Xu, Chenfeng, Peng, Chensheng, Zhao, Shuqi, Ho, Nhat, Pham, Alexander T., Ding, Mingyu, Tomizuka, Masayoshi, Zhan, Wei
Recent advancements have exploited diffusion models for the synthesis of either LiDAR point clouds or camera image data in driving scenarios. Despite their success in modeling single-modality data marginal distribution, there is an under-exploration
Externí odkaz:
http://arxiv.org/abs/2411.01123
Autor:
Jacobson, Philip, Xie, Yichen, Ding, Mingyu, Xu, Chenfeng, Tomizuka, Masayoshi, Zhan, Wei, Wu, Ming C.
Semi-supervised 3D object detection is a common strategy employed to circumvent the challenge of manually labeling large-scale autonomous driving perception datasets. Pseudo-labeling approaches to semi-supervised learning adopt a teacher-student fram
Externí odkaz:
http://arxiv.org/abs/2409.10901
Autor:
Wang, Yixiao, Tang, Chen, Sun, Lingfeng, Rossi, Simone, Xie, Yichen, Peng, Chensheng, Hannagan, Thomas, Sabatini, Stefano, Poerio, Nicola, Tomizuka, Masayoshi, Zhan, Wei
Diffusion models are promising for joint trajectory prediction and controllable generation in autonomous driving, but they face challenges of inefficient inference steps and high computational demands. To tackle these challenges, we introduce Optimal
Externí odkaz:
http://arxiv.org/abs/2408.00766
Autor:
Wang, Yixiao, Zhang, Yifei, Huo, Mingxiao, Tian, Ran, Zhang, Xiang, Xie, Yichen, Xu, Chenfeng, Ji, Pengliang, Zhan, Wei, Ding, Mingyu, Tomizuka, Masayoshi
The increasing complexity of tasks in robotics demands efficient strategies for multitask and continual learning. Traditional models typically rely on a universal policy for all tasks, facing challenges such as high computational costs and catastroph
Externí odkaz:
http://arxiv.org/abs/2407.01531
Road surface conditions, especially geometry profiles, enormously affect driving performance of autonomous vehicles. Vision-based online road reconstruction promisingly captures road information in advance. Existing solutions like monocular depth est
Externí odkaz:
http://arxiv.org/abs/2404.06605
The pretraining-finetuning paradigm has gained widespread adoption in vision tasks and other fields, yet it faces the significant challenge of high sample annotation costs. To mitigate this, the concept of active finetuning has emerged, aiming to sel
Externí odkaz:
http://arxiv.org/abs/2403.10069
End-to-end differentiable learning for autonomous driving (AD) has recently become a prominent paradigm. One main bottleneck lies in its voracious appetite for high-quality labeled data e.g. 3D bounding boxes and semantic segmentation, which are noto
Externí odkaz:
http://arxiv.org/abs/2403.02877
In the long-tailed recognition field, the Decoupled Training paradigm has demonstrated remarkable capabilities among various methods. This paradigm decouples the training process into separate representation learning and classifier re-training. Previ
Externí odkaz:
http://arxiv.org/abs/2403.00250
Autor:
Xie, Yichen, Chen, Hongge, Meyer, Gregory P., Lee, Yong Jae, Wolff, Eric M., Tomizuka, Masayoshi, Zhan, Wei, Chai, Yuning, Huang, Xin
Due to the lack of depth cues in images, multi-frame inputs are important for the success of vision-based perception, prediction, and planning in autonomous driving. Observations from different angles enable the recovery of 3D object states from 2D i
Externí odkaz:
http://arxiv.org/abs/2402.15583