Advancements in Shape Manipulation and 3D Content Generation

The field of robotic surface manipulation and 3D content generation is experiencing significant advancements, with a focus on developing innovative methods for shape control, texture generation, and part-level editing. Researchers are exploring the use of auxetic materials, latent diffusion models, and Gaussian processes to create more flexible and adaptable systems. Notable papers in this area include Reconfigurable Auxetic Devices, which demonstrates a novel approach to robotic surface manipulation using reconfigurable auxetic lattices, and SplatFont3D, which proposes a structure-aware text-to-3D artistic font generation framework. KeyPointDiffuser and GaussianBlender also show promising results in unsupervised 3D keypoint learning and instant stylization of 3D Gaussians, respectively.

Sources

Reconfigurable Auxetic Devices (RADs) for Robotic Surface Manipulation

SplatFont3D: Structure-Aware Text-to-3D Artistic Font Generation with Part-Level Style Control

Textured Word-As-Image illustration

PhyCustom: Towards Realistic Physical Customization in Text-to-Image Generation

FluxLab: Creating 3D Printable Shape-Changing Devices with Integrated Deformation Sensing

TEXTRIX: Latent Attribute Grid for Native Texture Generation and Beyond

KeyPointDiffuser: Unsupervised 3D Keypoint Learning via Latent Diffusion Models

GaussianBlender: Instant Stylization of 3D Gaussians with Disentangled Latent Spaces

SPLICE: Part-Level 3D Shape Editing from Local Semantic Extraction to Global Neural Mixing

Built with on top of