Thuật ngữ AI
Từ điển đầy đủ về Trí tuệ nhân tạo
Text-to-3D Generation
Generation of 3D objects from text descriptions using diffusion models combining text encoding and 3D geometric representations.
Image-to-3D Conversion
Process of volumetric or surface 3D reconstruction from 2D images using diffusion models to infer missing geometry.
Score-based 3D Modeling
Probabilistic framework using gradient scores to guide 3D shape generation through stochastic differential equations.
Tri-plane Representation
Compact encoding of 3D geometry in three orthogonal planes used by diffusion models for efficient generation of neural implicit fields.
3D-aware Diffusion
Diffusion models designed to understand and respect 3D consistency constraints when generating from multiple or single 2D views.
Geometry-aware Diffusion
Diffusion approach integrating explicit geometric constraints such as convexity, manifoldness, or symmetry for controlled 3D generation.
Surface Reconstruction Diffusion
Technique using diffusion to reconstruct smooth surfaces from partial or noisy data while preserving topological features.
Texture Diffusion 3D
Generation of consistent textures on 3D surfaces using diffusion models that respect UV continuity and view-dependent consistency.
Multi-view Diffusion
Synchronized diffusion process across multiple views of a 3D object to ensure geometric and textural consistency during reconstruction.
Sparse-view Reconstruction
3D reconstruction from a limited number of images using diffusion models to fill in missing information by learning 3D regularities.