no code implementations • 19 Mar 2024 • Quankai Gao, Qiangeng Xu, Zhe Cao, Ben Mildenhall, Wenchao Ma, Le Chen, Danhang Tang, Ulrich Neumann
While the optimization can draw photometric reference from the input videos or be regulated by generative models, directly supervising Gaussian motions remains underexplored.
1 code implementation • 24 Sep 2023 • Cho-Ying Wu, Quankai Gao, Chin-Cheng Hsu, Te-Lin Wu, Jing-Wen Chen, Ulrich Neumann
To facilitate our investigation for robustness and address limitations of previous works, we collect InSpaceType, a high-quality and high-resolution RGBD dataset for general indoor environments.
Indoor Monocular Depth Estimation Monocular Depth Estimation
1 code implementation • ICCV 2023 • Yiqi Zhong, Luming Liang, Ilya Zharkov, Ulrich Neumann
A central challenge of video prediction lies where the system has to reason the objects' future motions from image frames while simultaneously maintaining the consistency of their appearances across frames.
1 code implementation • ICCV 2023 • Quankai Gao, Qiangeng Xu, Hao Su, Ulrich Neumann, Zexiang Xu
In contrast to TensoRF which uses a global tensor and focuses on their vector-matrix decomposition, we propose to utilize a cloud of local tensors and apply the classic CANDECOMP/PARAFAC (CP) decomposition to factorize each tensor into triple vectors that express local feature distributions along spatial axes and compactly encode a local neural field.
no code implementations • 12 May 2023 • Cho-Ying Wu, Yiqi Zhong, Junying Wang, Ulrich Neumann
We instead propose fine-grained task that treats each RGB-D pair as a task in our meta-optimization.
Ranked #34 on Monocular Depth Estimation on NYU-Depth V2
no code implementations • CVPR 2023 • Junying Wang, Jae Shin Yoon, Tuanfeng Y. Wang, Krishna Kumar Singh, Ulrich Neumann
This paper presents a method to reconstruct a complete human geometry and texture from an image of a person with only partial body observed, e. g., a torso.
no code implementations • 20 Jul 2022 • Yiqi Zhong, Zhenyang Ni, Siheng Chen, Ulrich Neumann
In this work, we re-introduce this information as a new type of input data for trajectory forecasting systems: the local behavior data, which we conceptualize as a collection of location-specific historical trajectories.
no code implementations • 11 Jul 2022 • Bohan Tang, Yiqi Zhong, Chenxin Xu, Wei-Tao Wu, Ulrich Neumann, Yanfeng Wang, Ya zhang, Siheng Chen
Further, we apply the proposed framework to current SOTA multi-agent multi-modal forecasting systems as a plugin module, which enables the SOTA systems to 1) estimate the uncertainty in the multi-agent multi-modal trajectory forecasting task; 2) rank the multiple predictions and select the optimal one based on the estimated uncertainty.
1 code implementation • CVPR 2022 • Cho-Ying Wu, Chin-Cheng Hsu, Ulrich Neumann
This work digs into a root question in human perception: can face geometry be gleaned from one's voices?
Ranked #1 on 3D Face Modelling on Voxceleb-3D
1 code implementation • CVPR 2022 • Qiangeng Xu, Zexiang Xu, Julien Philip, Sai Bi, Zhixin Shu, Kalyan Sunkavalli, Ulrich Neumann
Point-NeRF combines the advantages of these two approaches by using neural 3D point clouds, with associated neural features, to model a radiance field.
2 code implementations • 4 Dec 2021 • Qiangeng Xu, Yiqi Zhong, Ulrich Neumann
Finally, the probability of occupancy is also integrated into a proposal refinement module to generate the final bounding boxes.
Ranked #2 on 3D Object Detection on KITTI Cars Moderate
1 code implementation • CVPR 2022 • Cho-Ying Wu, Jialiang Wang, Michael Hall, Ulrich Neumann, Shuochen Su
The majority of prior monocular depth estimation methods without groundtruth depth guidance focus on driving scenarios.
Ranked #1 on Monocular Depth Estimation on VA (Virtual Apartment)
no code implementations • NeurIPS 2021 • Bohan Tang, Yiqi Zhong, Ulrich Neumann, Gang Wang, Ya zhang, Siheng Chen
2) The results of trajectory forecasting benchmarks demonstrate that the CU-based framework steadily helps SOTA systems improve their performances.
4 code implementations • 19 Oct 2021 • Cho-Ying Wu, Qiangeng Xu, Ulrich Neumann
Our synergy process leverages a representation cycle for 3DMM parameters and 3D landmarks.
Ranked #1 on Face Alignment on AFLW
1 code implementation • 21 Apr 2021 • Cho-Ying Wu, Ke Xu, Chin-Cheng Hsu, Ulrich Neumann
This work focuses on the analysis that whether 3D face models can be learned from only the speech inputs of speakers.
1 code implementation • 16 Apr 2021 • Cho-Ying Wu, Qiangeng Xu, Ulrich Neumann
This work focuses on complete 3D facial geometry prediction, including 3D facial alignment via 3D face modeling and face orientation estimation using the proposed multi-task, multi-modal, and multi-representation landmark refinement network (M$^3$-LRN).
1 code implementation • 14 Jun 2020 • Cho-Ying Wu, Xiaoyan Hu, Michael Happold, Qiangeng Xu, Ulrich Neumann
Mask regression is based on 2D, 2. 5D, and 3D ROI using the pseudo-lidar and image-based representations.
Ranked #16 on Instance Segmentation on Cityscapes val (using extra training data)
1 code implementation • 15 Mar 2020 • Cho-Ying Wu, Ulrich Neumann
Recent sparse depth completion for lidars only focuses on the lower scenes and produces irregular estimations on the upper because existing datasets, such as KITTI, do not provide groundtruth for upper areas.
1 code implementation • CVPR 2020 • Qiangeng Xu, Xudong Sun, Cho-Ying Wu, Panqu Wang, Ulrich Neumann
Compared with popular sampling methods such as Farthest Point Sampling (FPS) and Ball Query, CAGQ achieves up to 50X speed-up.
1 code implementation • NeurIPS 2019 • Yiqi Zhong, Cho-Ying Wu, Suya You, Ulrich Neumann
Such a transformation enables CFCNet to predict features and reconstruct data of missing depth measurements according to their corresponding, transformed RGB features.
1 code implementation • 6 Jun 2019 • Cho-Ying Wu, Ulrich Neumann
Also, we propose to create building masks from semantic segmentation using an encoder-decoder network.
3 code implementations • NeurIPS 2019 • Qiangeng Xu, Weiyue Wang, Duygu Ceylan, Radomir Mech, Ulrich Neumann
Reconstructing 3D shapes from single-view images has been a long-standing research problem.
Ranked #1 on Single-View 3D Reconstruction on ShapeNetCore
1 code implementation • CVPR 2019 • Weiyue Wang, Duygu Ceylan, Radomir Mech, Ulrich Neumann
Given such a source 3D model and a target which can be a 2D image, 3D model, or a point cloud acquired as a depth scan, we introduce 3DN, an end-to-end network that deforms the source model to resemble the target.
no code implementations • 1 Nov 2018 • Cho Ying Wu, Ulrich Neumann
In this paper, we propose the multi-domain dictionary learn- ing (MDDL) to make dictionary learning-based classification more robust to data representing in different domains.
no code implementations • 1 Sep 2018 • Qiangeng Xu, Hanwang Zhang, Weiyue Wang, Peter N. Belhumeur, Ulrich Neumann
In this paper, we introduce a stochastic dynamics video infilling (SDVI) framework to generate frames between long intervals in a video.
4 code implementations • ECCV 2018 • Weiyue Wang, Ulrich Neumann
Convolutional neural networks (CNN) are limited by the lack of capability to handle geometric information due to the fixed grid kernel structure.
Ranked #6 on Semantic Segmentation on Stanford2D3D - RGBD
2 code implementations • CVPR 2018 • Qiangui Huang, Weiyue Wang, Ulrich Neumann
The key component of the RSNet is a lightweight local dependency module.
Ranked #46 on Semantic Segmentation on S3DIS
no code implementations • 23 Jan 2018 • Qiangui Huang, Kevin Zhou, Suya You, Ulrich Neumann
Specifically, we introduce a "try-and-learn" algorithm to train pruning agents that remove unnecessary CNN filters in a data-driven way.
1 code implementation • CVPR 2018 • Weiyue Wang, Ronald Yu, Qiangui Huang, Ulrich Neumann
Experimental results on various 3D scenes show the effectiveness of our method on 3D instance segmentation, and we also evaluate the capability of SGPN to improve 3D object detection and semantic segmentation results.
Ranked #1 on 3D Semantic Instance Segmentation on ScanNetV1
1 code implementation • ICCV 2017 • Weiyue Wang, Qiangui Huang, Suya You, Chao Yang, Ulrich Neumann
The 3D-ED-GAN is a 3D convolutional neural network trained with a generative adversarial paradigm to fill missing 3D data in low-resolution.
no code implementations • 22 Nov 2016 • Qiangui Huang, Weiyue Wang, Kevin Zhou, Suya You, Ulrich Neumann
A novel neural network architecture is built for scene labeling tasks where one of the variants of the new RNN unit, Gated Recurrent Unit with Explicit Long-range Conditioning (GRU-ELC), is used to model multi scale contextual dependencies in images.