https://github.com/ganlab/ostra
공식 코드A One Stop 3D Target Reconstruction and multilevel Segmentation Method
태스크의 하위 분류와 논문·코드·평가·기법·데이터셋을 탐색하세요.
Image: Choy et al
관련 논문과 평가 결과에 연결된 구현입니다. 출처에서 확인된 저자 구현에는 공식 코드 표시가 있습니다.
A One Stop 3D Target Reconstruction and multilevel Segmentation Method
UMIFormer: Mining the Correlations between Similar Tokens for Multi-View 3D Reconstruction
Modeling 3D Surface Manifolds with a Locally Conditioned Atlas
ComplexGen: CAD Reconstruction by B-Rep Chain Complex Generation
Pix2Vox: Context-aware 3D Reconstruction from Single and Multi-view Images
Next-best-view Regression using a 3D Convolutional Neural Network
Learning Free-Form Deformations for 3D Object Reconstruction
Implicit Functions in Feature Space for 3D Shape Reconstruction and Completion
Self-calibrating Photometric Stereo by Neural Inverse Rendering
LeanGaussian: Breaking Pixel or Point Cloud Correspondence in Modeling 3D Gaussians
Robot See Robot Do: Imitating Articulated Object Manipulation with Monocular 4D Reconstruction
ExtrudeNet: Unsupervised Inverse Sketch-and-Extrude for Shape Parsing
Grasping Field: Learning Implicit Representations for Human Grasps
Grasping Field: Learning Implicit Representations for Human Grasps
DmifNet:3D Shape Reconstruction Based on Dynamic Multi-Branch Information Fusion
Long-Range Grouping Transformer for Multi-View 3D Reconstruction
Neural Face Identification in a 2D Wireframe Projection of a Manifold Object
Motion Blender Gaussian Splatting for Dynamic Scene Reconstruction
Share With Thy Neighbors: Single-View Reconstruction by Cross-Instance Consistency
GaussianGrasper: 3D Language Gaussian Splatting for Open-vocabulary Robotic Grasping
Spacetime Gaussian Feature Splatting for Real-Time Dynamic View Synthesis
NU-MCC: Multiview Compressive Coding with Neighborhood Decoder and Repulsive UDF
Hierarchical Neural Coding for Controllable CAD Model Generation
Deep Predictive Motion Tracking in Magnetic Resonance Imaging: Application to Fetal Imaging
Adding Knowledge to Unsupervised Algorithms for the Recognition of Intent
Feature Splatting for Better Novel View Synthesis with Low Overlap
NTO3D: Neural Target Object 3D Reconstruction with Segment Anything
From Image Collections to Point Clouds with Self-supervised Shape and Pose Networks
출처: Papers with Code · CC BY-SA 4.0. 아카이브 자료와 새로 수집한 논문은 별도로 관리합니다.