Results 11 to 20 of about 4,051,827 (361)
VoxFormer: Sparse Voxel Transformer for Camera-Based 3D Semantic Scene Completion [PDF]
Computer Vision and Pattern Recognition, 2023 Humans can easily imagine the complete 3D geometry of occluded objects and scenes. This appealing ability is vital for recognition and understanding.
Yiming Li +7 more
semanticscholar +1 more source
RepoCoder: Repository-Level Code Completion Through Iterative Retrieval and Generation [PDF]
Conference on Empirical Methods in Natural Language Processing, 2023 The task of repository-level code completion is to continue writing the unfinished code based on a broader context of the repository. While for automated code completion tools, it is difficult to utilize the useful information scattered in different ...
Fengji Zhang +7 more
semanticscholar +1 more source
AutoSDF: Shape Priors for 3D Completion, Reconstruction and Generation [PDF]
Computer Vision and Pattern Recognition, 2022 Powerful priors allow us to perform inference with in-sufficient information. In this paper, we propose an au-toregressive prior for 3D shapes to solve multimodal 3D tasks such as shape completion, reconstruction, and gener-ation.
Paritosh Mittal +3 more
semanticscholar +1 more source
CompletionFormer: Depth Completion with Convolutions and Vision Transformers [PDF]
Computer Vision and Pattern Recognition, 2023 Given sparse depths and the corresponding RGB images, depth completion aims at spatially propagating the sparse measurements throughout the whole image to get a dense depth prediction.
Youming Zhang +5 more
semanticscholar +1 more source
PoinTr: Diverse Point Cloud Completion with Geometry-Aware Transformers [PDF]
IEEE International Conference on Computer Vision, 2021 Point clouds captured in real-world applications are of-ten incomplete due to the limited sensor resolution, single viewpoint, and occlusion. Therefore, recovering the complete point clouds from partial ones becomes an indispensable task in many ...
Xumin Yu +5 more
semanticscholar +1 more source
RepoBench: Benchmarking Repository-Level Code Auto-Completion Systems [PDF]
International Conference on Learning Representations, 2023 Large Language Models (LLMs) have greatly advanced code auto-completion systems, with a potential for substantial productivity enhancements for developers.
Tianyang Liu, Canwen Xu, Julian McAuley
semanticscholar +1 more source
3D Shape Generation and Completion through Point-Voxel Diffusion [PDF]
IEEE International Conference on Computer Vision, 2021 We propose a novel approach for probabilistic generative modeling of 3D shapes. Unlike most existing models that learn to deterministically translate a latent vector to a shape, our model, Point-Voxel Diffusion (PVD), is a unified, probabilistic ...
Linqi Zhou, Yilun Du, Jiajun Wu
semanticscholar +1 more source
MonoScene: Monocular 3D Semantic Scene Completion [PDF]
Computer Vision and Pattern Recognition, 2021 MonoScene proposes a 3D Semantic Scene Completion (SSC) framework, where the dense geometry and semantics of a scene are inferred from a single monocular RGB image.
Anh-Quan Cao, Raoul de Charette
semanticscholar +1 more source
SDFusion: Multimodal 3D Shape Completion, Reconstruction, and Generation [PDF]
Computer Vision and Pattern Recognition, 2022 In this work, we present a novel framework built to sim-plify 3D asset generation for amateur users. To enable interactive generation, our method supports a variety of input modalities that can be easily provided by a human, in-cluding images, text ...
Yen-Chi Cheng +4 more
semanticscholar +1 more source
PENet: Towards Precise and Efficient Image Guided Depth Completion [PDF]
IEEE International Conference on Robotics and Automation, 2021 Image guided depth completion is the task of generating a dense depth map from a sparse depth map and a high quality image. In this task, how to fuse the color and depth modalities plays an important role in achieving good performance.
Mu Hu +5 more
semanticscholar +1 more source