Zobrazeno 1 - 10
of 132
pro vyhledávání: '"Wang, Hengyi"'
Autor:
Wang, Hengyi, Agapito, Lourdes
We present Spann3R, a novel approach for dense 3D reconstruction from ordered or unordered image collections. Built on the DUSt3R paradigm, Spann3R uses a transformer-based architecture to directly regress pointmaps from images without any prior know
Externí odkaz:
http://arxiv.org/abs/2408.16061
Vision transformers (ViTs) have emerged as a significant area of focus, particularly for their capacity to be jointly trained with large language models and to serve as robust vision foundation models. Yet, the development of trustworthy explanation
Externí odkaz:
http://arxiv.org/abs/2406.12649
Autor:
Wang, Hengyi, Shi, Haizhou, Tan, Shiwei, Qin, Weiyi, Wang, Wenyuan, Zhang, Tunyu, Nambi, Akshay, Ganu, Tanuja, Wang, Hao
Multimodal Large Language Models (MLLMs) have shown significant promise in various applications, leading to broad interest from researchers and practitioners alike. However, a comprehensive evaluation of their long-context capabilities remains undere
Externí odkaz:
http://arxiv.org/abs/2406.11230
Autor:
Shi, Haizhou, Xu, Zihao, Wang, Hengyi, Qin, Weiyi, Wang, Wenyuan, Wang, Yibin, Wang, Zifeng, Ebrahimi, Sayna, Wang, Hao
The recent success of large language models (LLMs) trained on static, pre-collected, general datasets has sparked numerous research directions and applications. One such direction addresses the non-trivial challenge of integrating pre-trained LLMs in
Externí odkaz:
http://arxiv.org/abs/2404.16789
Autor:
Hu, Naihong, Wang, Hengyi
The centre of two-parameter quantum groups $U_{r,s}(\mathfrak{g})$ is determined through the Harish-Chandra homomorphism. Based on the Rosso form and the representation theory of weight modules, we prove that when rank $\mathfrak{g}$ is even, the Har
Externí odkaz:
http://arxiv.org/abs/2402.12793
Neural rendering has demonstrated remarkable success in dynamic scene reconstruction. Thanks to the expressiveness of neural representations, prior works can accurately capture the motion and achieve high-fidelity reconstruction of the target object.
Externí odkaz:
http://arxiv.org/abs/2312.00778
We present Co-SLAM, a neural RGB-D SLAM system based on a hybrid representation, that performs robust camera tracking and high-fidelity surface reconstruction in real time. Co-SLAM represents the scene as a multi-resolution hash-grid to exploit its h
Externí odkaz:
http://arxiv.org/abs/2304.14377
Autor:
Wang, Hengyi, Oh, Changjae
We present a refinement framework to boost the performance of pre-trained semi-supervised video object segmentation (VOS) models. Our work is based on scale inconsistency, which is motivated by the observation that existing VOS models generate incons
Externí odkaz:
http://arxiv.org/abs/2205.01197
We present methods to estimate the physical properties of household containers and their fillings manipulated by humans. We use a lightweight, pre-trained convolutional neural network with coordinate attention as a backbone model of the pipelines to
Externí odkaz:
http://arxiv.org/abs/2203.01192
Autor:
Wang, Lianxin, Meng, Jinhui, Yu, Xiaomiao, Wang, Jie, Zhang, Yuying, Zhang, Man, Zhang, Yuxi, Wang, Hengyi, Feng, Huawei, Tian, Qifeng, Zhang, Li, Liu, Hongsheng
Publikováno v:
In Archives of Biochemistry and Biophysics October 2024 760