Advances in Human Motion Synthesis and 3D Scene Generation

The field of human motion synthesis and 3D scene generation is rapidly advancing, with a focus on developing more realistic and diverse models. Recent research has explored the use of deterministic-to-stochastic latent feature mapping, training-free scene-aware text-to-motion generation, and generalist models that can handle multiple tasks within a single framework. These innovations have improved the accuracy and efficiency of human motion synthesis, enabling the generation of more realistic and diverse motion sequences. Additionally, advances in 3D scene generation have led to the development of more realistic and interactive environments, with applications in fields such as gaming, virtual reality, and urban planning. Notable papers include GENMO, which presents a unified generalist model for human motion estimation and generation, and Scenethesis, which introduces a training-free agentic framework for 3D scene generation. Overall, these developments are poised to have a significant impact on the field, enabling the creation of more realistic and engaging virtual environments.

Sources

Deterministic-to-Stochastic Diverse Latent Feature Mapping for Human Motion Synthesis

TSTMotion: Training-free Scene-awarenText-to-motion Generation

GENMO: A GENeralist Model for Human MOtion

FreeInsert: Disentangled Text-Guided Object Insertion in 3D Gaussian Scene without Spatial Priors

Real-time Spatial Retrieval Augmented Generation for Urban Environments

Scenethesis: A Language and Vision Agentic Framework for 3D Scene Generation

StableMotion: Training Motion Cleanup Models with Unpaired Corrupted Data

STORY2GAME: Generating (Almost) Everything in an Interactive Fiction Game

BuildingBlock: A Hybrid Approach for Structured Building Generation

ReAlign: Bilingual Text-to-Motion Generation via Step-Aware Reward-Guided Alignment

PlaceIt3D: Language-Guided Object Placement in Real 3D Scenes

Generating Physically Stable and Buildable LEGO Designs from Text

Built with on top of