Content & Design

motion-diffusion-model is an open-source PyTorch implementation of the "Human Motion Diffusion Model" paper, designed for generating and editing human motion sequences. The tool boasts significant speed improvements, now running 40X faster with a 50-diffusion-step model and optimized CLIP calling. It supports various tasks including text-to-motion, action-to-motion, and unconstrained motion synthesis. Users can generate motions from text prompts or actions, render SMPL meshes, and perform motion editing such as in-between and upper-body modifications. The project also integrates DiP for ultra-fast text-to-motion and offers features like DistilBERT text encoder support and dataset caching for faster loading.

DiscoFaceGAN is a TensorFlow-based implementation for disentangled and controllable face image generation, as presented in a CVPR 2020 Oral paper. This tool allows for the creation of virtual people's faces with precise control over identity, expression, pose, and illumination. It achieves this through 3D imitative-contrastive learning, embedding 3D priors into adversarial learning to imitate the image formation of a 3D face deformation and rendering process. A key feature is its factor disentanglement, ensuring that changing one factor (e.g., expression) does not affect others. The tool also supports reference-based generation, real image pose manipulation, lighting editing, and expression transfer, making it valuable for researchers and developers working with facial image synthesis and manipulation.

motion-latent-diffusion (MLD) is an open-source text-to-motion and action-to-motion diffusion model that significantly improves motion quality and speed compared to previous models. It operates by executing commands via motion diffusion in latent space, offering a state-of-the-art solution for generating realistic and fluid movements. The tool is designed for researchers and developers working with 3D animation and motion synthesis, providing scripts for quick setup, pre-trained models, and detailed guidance for training custom models. MLD also includes features for visualizing latent space and rendering SMPL meshes, making it a comprehensive platform for advanced motion diffusion tasks.

AIgentor is a public AI character chat platform designed for fast, no-login access to AI conversations. Users can browse a wide array of characters, start chatting instantly, and even create their own public characters anonymously without needing a traditional account. The platform prioritizes a low-friction experience, offering cleaner navigation and faster loading times compared to other character chat sites. It is built around free access, supported by advertising, and aims to make character chat feel immediate by minimizing unnecessary steps or forced account systems. Character creation is reviewed and moderated to ensure content suitability and safety.

Bevelify.com is a domain name currently listed for sale on HugeDomains.com. The platform facilitates the purchase of domain names, offering a secure transaction process and a 30-day money-back guarantee for customer satisfaction. Buyers can choose between an outright purchase or a payment plan spread over 24 months with 0% interest. HugeDomains.com also provides quick delivery of the domain, typically within one to two hours, and offers support for transferring the domain to other registrars like GoDaddy. Additional services like email packages and privacy protection are available through their registrar, NameBright.com.

Opalai Scanto3D is a spatial AI tool designed to simplify the creation of 2D and 3D CAD models and Building Information Models (BIM). Users can leverage the ScanTo3D mobile app to scan environments on the go, eliminating the need for bulky LiDAR scanners. The platform also allows for the transformation of existing LiDAR or 360 camera data into detailed BIMs in three simple steps. Beyond 3D modeling, OpalAi offers a suite of spatial AI tools including FireVision for wildfire intelligence, Urbanomy for smarter city streets, and Astro Adjuster for accelerating video analysis with Vision Language Models. It caters to various industries such as PropTech, InsurTech, Space Tech, and Transport Tech, providing solutions for property modeling, claim assessments, satellite imagery analysis, and infrastructure management.

3dmm_cnn is an open-source project that provides an end-to-end solution for regressing robust and discriminative 3D Morphable Models (3DMM) from unconstrained 2D face images. Utilizing a very deep neural network, this tool estimates the 3D facial shape and texture, outputting a standard PLY file. It is designed to work effectively even with challenging datasets like LFW, YTF, and IJB-A benchmarks. A key feature is its speed, as it avoids expensive, iterative optimization loops for shape regression. Additionally, 3dmm_cnn includes extra code for head pose and expression estimation using detected facial landmarks and the regressed 3D face model. The project is primarily developed in C++ and Python, with dependencies including Dlib, OpenCV, Caffe, and Numpy, and has been tested on Linux.

Xspiral is an online 3D visualization tool designed for creating and interacting with 3D models. It integrates both 2D and 3D design capabilities, enhanced with AI features to streamline the creative process. The platform supports a variety of design and visualization tasks, making it suitable for users who require an accessible online solution for 3D design. While specific features are not detailed on the provided pages, the tool aims to offer a comprehensive environment for 3D creation and manipulation, catering to diverse project needs in the realm of digital design and animation.

HelloAR is an industry-leading platform designed to elevate brand engagement through patented AI-enhanced 3D visuals and immersive 360° experiences. It offers specialized solutions like Hello Ramp for automotive sales, providing AI Catalog Transformations and streamlined inventory management. Hello Shape allows users to craft compelling interactive 3D and augmented reality catalogs, bringing products to life with interactive and immersive showcases. Furthermore, Hello Shift enables the creation of virtual experiences with multiple templates, fostering in-store-like engagement. The platform boasts impressive metrics, including 100% consistent images, 60% CTR growth, 80% faster launches, and 75% cost reduction, making it a powerful tool for brands looking to enhance their visual marketing and sales strategies.

Google Earth Studio is a web-based animation tool designed for creating professional animations using Google Earth’s satellite and 3D imagery. It offers powerful motion design capabilities directly in the browser, utilizing keyframe animation similar to industry-standard tools. Users can easily move the globe, set keyframes, and animate custom attributes like the sun's position or camera's field of view. The tool also provides quick-start projects with templates for orbits or point-to-point flights, making it accessible even without prior animation experience. Additionally, it supports 3D camera export to Adobe After Effects, enabling users to add map labels and pins in post-production. Earth Studio is free to use and works with the desktop version of Google Chrome.

AnimeShorts AI is an innovative platform designed for generating high-quality AI anime art, animation videos, and comics. Users can easily create character-stable AI anime art, ensuring consistency across their projects. The tool supports the creation of AI comics and webtoons, making it a versatile solution for various animation and illustration needs. It offers AI tools, models, templates, and character options to streamline the creative process, enabling users to unlock their artistic potential in the anime domain.

VeggieAI.dance is an innovative AI dance generator that allows users to create dynamic dance videos by simply uploading photos. Serving as a Viggle AI alternative, it specializes in transforming static images into animated dances, providing an engaging and interactive experience. The platform offers advanced 3D character animations from text prompts, ensuring smooth and lifelike movements grounded in realistic physics. Users can customize backgrounds to personalize their dance videos and utilize the "Veggie AI head top" feature to sync head movements with music. VeggieAI.dance provides free online access with a free quota, prioritizing data security for all uploaded content.

MVSplat is an open-source project presented at ECCV 2024, designed for efficient 3D Gaussian Splatting from sparse multi-view images. It provides a robust framework for generating 3D models from a limited set of input images, making it valuable for computer vision research and 3D graphics development. The tool supports evaluation, video rendering, and training with various datasets like RealEstate10K, ACID, and DTU. It also offers features for fine-tuning from released weights and conducting ablations to understand model components. MVSplat is built on PyTorch Lightning, allowing for scalable training on multiple GPUs and nodes, and has extensions like PanSplat for higher resolutions and MVSplat360 for 360° NVS from fewer views.

SAM 3D is an online tool powered by Meta’s open-source SAM 3D research models, enabling users to transform single images into detailed, realistic 3D assets. It supports both objects and human bodies, offering specialized architectures like SAM 3D Objects for scene-aware reconstruction and SAM 3D Body for precise human digitization using the Meta Momentum Human Rig (MHR). The tool boasts unmatched robustness in real-world environments, handling low light, heavy occlusion, and non-standard camera angles with high fidelity. Users can upload an image, prompt the model by clicking on the desired object or person, and instantly generate a high-quality 3D mesh. Assets can be exported in standard formats such as .OBJ, .GLB, or MHR, ready for integration into various 3D software like Blender, Unity, or Unreal Engine. It is designed for fast inference without requiring local setup or a powerful GPU.

MeshGPT is an advanced AI tool focused on generating triangle meshes through the application of decoder-only transformers. This innovative approach allows for the creation of 3D models from various inputs, providing a powerful solution for complex geometric structures. The tool is primarily aimed at AI researchers and graphics programmers who are exploring the frontiers of 3D model generation and computational geometry. Its core functionality revolves around transforming data into detailed mesh representations, which are fundamental in 3D graphics, simulations, and virtual reality applications. The official code release is available on GitHub, indicating its open-source nature and accessibility for technical users to experiment and build upon.

Animate & Create is a Pittsburgh-based creative and development studio specializing in premium 3D animation, CGI product visuals, FOOH campaigns, and AI-assisted animation. They provide full-service creative production for modern brands, helping e-commerce teams, agencies, founders, and various businesses transform product ideas into high-quality motion assets. Their core services include 3D product animation, product visualization, CGI commercials, motion graphics, AI animation, FOOH advertising, web development, and app experiences. The studio blends strategy, design, animation, development, and fulfillment workflows to deliver visuals that are both impactful and performance-driven. Projects can be handled as one-time deliverables or through recurring credit-based production plans.

The 3 Pixel Grid is a free and open-source project offering a collection of tiny 3x3 animated indicators, designed for micro-interactions within web and application interfaces. It provides a simple and efficient way to add subtle animations, such as loaders or status indicators, to any project. Users can easily browse through various pre-designed patterns, click to copy the configuration code, and then drop it directly into their development environment. This tool is ideal for developers and designers looking to enhance user experience with lightweight, customizable animations without the need for complex animation software or extensive coding.

3D House Planner is a professional web-based application designed for creating detailed 3D home designs. It enables users to build apartments and houses with ease, offering intuitive controls for sketching 3D models, constructing walls, floors, roofs, and frames. The platform supports importing various 3D file formats including GLB, GLTF, OBJ, and Babylon models, and allows users to export their designs. A standout feature is its AI tool, which can generate 3D floor plans from imported floor plan images with a single click. Users can also browse a catalog of thousands of objects to furnish and decorate their homes, design up to 50 floors, and customize designs with a wide range of materials and textures, including Physically Based Rendering (PBR) for realistic visuals. The tool emphasizes privacy, with no login required and designs saved locally on the user's device.

Robotic World Model (RWM) is a GitHub repository that extends Isaac Lab, a robotics simulation platform, with environments and training pipelines for advanced model-based reinforcement learning. It specifically supports Robotic World Model (RWM) and Uncertainty-Aware Robotic World Model (RWM-U) methods. The tool enables joint training of policies and neural dynamics models within Isaac Lab, as well as offline policy training using learned neural network dynamics without requiring a simulator. Users can evaluate model-based versus model-free policies, visualize autoregressive imagination rollouts from learned dynamics, and observe trained policies in Isaac Lab. It offers configurations for model inputs/outputs, architecture, and training hyperparameters, making it a comprehensive solution for researchers and developers in robotics.

trt_pose is a powerful open-source project designed for real-time pose estimation, leveraging the acceleration capabilities of NVIDIA TensorRT. It is specifically optimized for NVIDIA Jetson platforms, making it ideal for edge AI applications. The project includes pre-trained models for human pose estimation, allowing users to easily detect key features like left eye, left elbow, and right ankle. Beyond human pose, trt_pose offers training scripts to train on any keypoint task data in MSCOCO format, providing flexibility for various keypoint detection tasks. It also integrates with other NVIDIA AI-IOT projects like trt_pose_hand for hand pose estimation and torch2trt for PyTorch to TensorRT conversion, forming a comprehensive ecosystem for AI development on NVIDIA hardware.

Viziverse AI delivers a groundbreaking headsetless VR experience, leveraging Vision AI to facilitate real-time 3D interaction across various devices. This innovative technology allows players to engage with games in a new way and empowers developers to create immersive content without the need for traditional VR headsets. The patented Reflective Reality™ platform eliminates the cost, logistics, and safety concerns associated with head-mounted hardware, making spatial computing accessible with zero hardware friction. Viziverse AI powers next-generation experiences in gaming, educational and training programs, experiential marketing, fitness, digital health, and 3D design, offering a scalable solution for diverse industries.

Tourly Immersive Audio Guide is a cutting-edge mobile application designed to enhance your travel experiences by transforming local landmarks into captivating, audio-guided tours. Leveraging augmented reality and AI-processed descriptions, the app instantly provides rich, geo-referenced articles about nearby points of interest. This innovative approach allows users to discover the world around them in an immersive way, making each location's history and significance come alive through audio. It's an ideal companion for travelers seeking a deeper understanding of their surroundings without the need for traditional tour guides, offering a seamless and informative exploration.

3DGS-to-PC provides scripts to convert 3D Gaussian Splatting scenes into dense point clouds, and optionally meshes. This tool addresses the compatibility issue of specialized Gaussian renders by transforming them into widely supported point cloud formats. Users can customize the point cloud creation with various arguments, including input/output paths, renderer type (CUDA or Python), number of points, and visibility thresholds. It also offers features for noise reduction, such as culling Gaussians based on contribution or distance from predicted surfaces, and statistical outlier removal. The tool supports generating meshes with adjustable Poisson depth and Laplacian smoothing iterations, making it versatile for different 3D handling software.

Disney Research is a network of research laboratories within The Walt Disney Company, dedicated to pushing the boundaries of technological innovation. The organization conducts in-depth research across a broad spectrum of fields, including computer graphics, advanced video processing techniques, and robotics. By engaging with the global research community, Disney Research actively contributes to scientific knowledge through publications and collaborations. Its work aims to develop new technologies and creative applications that can enhance various aspects of entertainment and beyond, reflecting Disney's commitment to innovation and future-forward thinking.