3D AI Tools

InterMask can generate high-quality 3D human interactions from text descriptions. It captures complex movements between two people while also allowing for reaction generation without changing the model.

TreeMeshGPT can generate detailed 3D meshes from point clouds using Autoregressive Tree Sequencing. This technique allows for better mesh detail and achieves a 22% reduction in data size during processing.

DART can generate high-quality human motions in real-time, achieving over 300 frames per second on a single RTX 4090 GPU. It combines text inputs with spatial constraints, allowing for tasks like reaching waypoints and interacting with scenes.

Make-It-Animatable can auto-rig any 3D humanoid model for animation in under one second. It generates high-quality blend weights and bones, and works with various 3D formats, ensuring accuracy even for non-standard skeletons.

StdGEN can generate high-quality 3D characters from a single image in just three minutes. It breaks down characters into parts like body, clothes, and hair, using a transformer-based model for great results in 3D anime character generation.

So far it has been tough to imagine the benefits of AI agents. Most of what we’ve seen from that domain has been focused on NPC simulations or solving text-based goals. 3D-GPT is a new framework that utilizes LLMs for instruction-driven 3D modeling by breaking down 3D modeling tasks into manageable segments to procedurally generate 3D scenes. I recently started to dig into Blender and I pray this gets open sourced at one point.

Phidias can generate high-quality 3D assets from text, images, and 3D references. It uses a method called reference-augmented diffusion to improve quality and speed, achieving results in just a few seconds.

EventEgo3D++ can capture 3D human motion using a monocular event camera with a fisheye lens. It works well in low-light and high-speed conditions, providing real-time 3D pose updates at 140Hz with high accuracy compared to RGB-based methods.

Cyberpunk brain dances are becoming a thing! D-NPC can turn videos into dynamic neural point clouds aka 4D scenes which makes it possible to watch a scene from another perspective.

FreeTimeGS can reconstruct dynamic 3D scenes in real-time using Gaussian primitives that can appear at different times and places.

MagicArticulate can rig static 3D models and make them ready for animation. Works on both humanoid and non-humanoid objects.

MEGASAM can estimate camera parameters and depth maps from casual monocular videos.

Cycle3D can generate high-quality and consistent 3D content from a single unposed image. This approach enhances texture consistency and multi-view coherence, significantly improving the quality of the final 3D reconstruction.

LIFe-GoM can create animatable 3D human avatars from sparse multi view images in under 1 second. It renders high-quality images at 95.1 frames per second.

DressRecon can create 3D human body models from single videos. It handles loose clothing and objects well, achieving high-quality results by combining general human shapes with specific video movements.

Google DeepMind has been researching 4DiM, a cascaded diffusion model for 4D novel view synthesis. It can generate 3D scenes with temporal dynamics from a single image and a set of camera poses and timestamps.

Dora can generated 3D assets from images which are ready for diffusion-based character control in modern 3D engines, such as Unity 3D, in real-time.

LayerPano3D can generate immersive 3D scenes from a single text prompt by breaking a 2D panorama into depth layers.

TeSMo is a method for text-controlled scene-aware motion generation and is able to generate realistic and diverse human-object interactions, such as navigation and sitting, in different scenes with various object shapes, orientations, initial body positions, and poses.

MotionLab can generate and edit human motion and supports text-based and trajectory-based motion creation.