Content & Design

Looq AI offers a comprehensive platform for spatial intelligence, designed to improve bottom-line costs and top-line revenues in the built world. It provides precision modeling, insights, and seamless data integration. The platform simplifies data capture using handheld photogrammetry, processes data with AI-enabled technology to generate survey-grade 3D point clouds and 2D data, and allows for easy collaboration and sharing of insights. Looq AI is particularly beneficial for survey/civil engineering, T&D engineering, and utilities, enabling users to quickly and safely capture real-world conditions, analyze accurate models, and deliver actionable data for various engineering and GIS workflows.

Glyf Journal serves as a platform for short reflections and everyday insights, featuring articles across diverse subjects. Currently, a significant portion of its content focuses on providing detailed information and comparisons of LLC formation services in various US states, such as Illinois, Minnesota, Nebraska, North Dakota, Nevada, Utah, Washington, Virginia, and Montana. The journal aims to offer resources and guidance for individuals looking to establish Limited Liability Companies, covering benefits, processes, and service evaluations. While the meta tags suggest a broader scope of 'content, articles, information, resources,' the live website content is heavily concentrated on business formation topics.

GauStudio is a modular framework designed to support and accelerate research and development in the rapidly advancing field of 3D Gaussian Splatting (3DGS) and its diverse applications. It offers functionalities like mesh extraction and rendering, and supports various 3DGS methods. The framework includes curated datasets for evaluating 3DGS methods under diverse conditions, including synthetic datasets and real-world scenes with high-quality normal annotations. GauStudio also provides LoFTR-based initial point clouds for better initialization and plans to release more 3DGS-based methods, dataset loaders, and visualization tools in the near future. It is released under the MIT License, with commercial cooperation welcomed.

Learnable-triangulation-pytorch is an official PyTorch implementation of the paper "Learnable Triangulation of Human Pose" (ICCV 2019, oral). This open-source project focuses on 3D human pose estimation from multiple cameras, offering two novel methods: Algebraic and Volumetric learnable triangulation. These methods significantly outperform previous state-of-the-art techniques, with the Volumetric model achieving a 2.4 times reduction in error. The repository provides code for training and evaluation, supports both single and multi-GPU setups, and includes pretrained models and configurations for the Human3.6M dataset. It is designed for researchers and engineers working on advanced computer vision tasks, particularly in human pose estimation.

FLUX Animation Creator is an innovative AI tool hosted on Hugging Face Spaces, designed to simplify the creation of animated GIFs. By simply entering a text prompt, users can describe the animation they envision, and the application will generate a series of images that seamlessly animate into a GIF. This user-friendly platform makes it accessible for anyone to transform their ideas into dynamic visual content without needing complex animation software or technical skills. It's an ideal solution for quickly producing engaging animated content for various purposes.

OmniSVG is the first family of end-to-end multimodal SVG generators, leveraging advanced pre-trained Vision-Language Models (VLMs) to create highly complex and detailed Scalable Vector Graphics. This innovative tool can generate a wide range of visuals, from basic icons to elaborate anime characters. It introduces MMSVG-2M, a multimodal dataset featuring two million richly annotated SVG assets, alongside a standardized evaluation protocol for conditional SVG generation tasks. OmniSVG offers various model weights, including OmniSVG1.1_8B and OmniSVG1.1_4B, and provides both text-to-SVG and image-to-SVG generation capabilities. An interactive demo is available via Gradio, allowing users to experiment with its generation features.

Gen6D is an open-source project focused on generalizable model-free 6-DoF object pose estimation from RGB images. Developed for ECCV 2022, this tool allows users to estimate the 6-DoF poses of previously unseen objects. It comes with pretrained models and evaluation codes, enabling immediate use for various tasks. The project supports pose estimation on custom objects and provides comprehensive training codes for users who wish to fine-tune or train their own models. Key features include detection, viewpoint selection, and pose refinement, with intermediate and final qualitative results saved for analysis. The repository also details the process for creating GenMOP objects for evaluation and acknowledges contributions from several other open-source projects and datasets.

GaussianSplats3D offers a Three.js-based implementation for 3D Gaussian splatting, enabling the generation and real-time rendering of 3D scenes from 2D images directly in a web browser. Unlike CUDA-based solutions requiring native machine execution, this project focuses on web accessibility. It supports various 3D scene formats including .ply, standard .splat, and its custom compressed .ksplat files, which can be converted for optimal performance. The tool is designed for developers and 3D enthusiasts, providing organized ES modules, built-in WebXR support, and optimizations like octree culling and WASM SIMD-based splat sorting. While no longer in active development, it serves as a foundational open-source project for web-based 3D Gaussian splatting.

SAX-NeRF is a comprehensive, open-source toolbox designed for X-ray 3D reconstruction, encompassing both novel view synthesis (NVS) and computed tomography (CT) reconstruction. This powerful library supports 11 state-of-the-art methods, including six NeRF-based, two 3DGS-based, two optimization-based, and one analytical method. It provides researchers and developers with tools for tasks such as generating turntable videos and synthesizing data. The project emphasizes structure-aware sparse-view X-ray 3D reconstruction and has been recognized at CVPR 2024. It offers code, models, and training logs, making it a valuable resource for advancing medical imaging and related applications.

LERF (Language Embedded Radiance Fields) is an open-source implementation designed for language-guided 3D scene reconstruction. It integrates seamlessly with the Nerfstudio framework, providing a platform for researchers and developers to experiment with neural radiance fields. Users can launch training with specified data folders and connect to a viewer to visualize relevancy maps based on text input. The tool offers different implementations like `lerf-big` for larger models and `lerf-lite` for reduced memory footprint, catering to various GPU capacities. It also supports custom image encoders, allowing for modular extension and integration of different models like OpenAI's CLIP or OpenCLIP. LERF is particularly useful for those working on advanced computer vision and 3D reconstruction projects.

LongSplat is an advanced framework designed for robust 3D Gaussian Splatting from casually captured long videos. It addresses the challenges of unposed video data through incremental joint optimization, which includes pose estimation with MASt3R and adaptive octree anchoring. This allows for high-quality novel view synthesis from free viewpoints, making it suitable for various applications in computer vision and 3D graphics. The tool is memory-efficient and scalable, capable of handling varied resolutions and frame rates. It also provides a conversion script to transform its results into the standard 3DGS format for compatibility with general 3DGS viewers.

Ripl is a high-performance canvas rendering library designed to provide a rich set of features through a simple and familiar API. It offers a unified API for drawing abstract shapes to various contexts, making it versatile for different development needs. The library emphasizes simplicity, modularity, and performance, allowing developers to include only the necessary components. Ripl emulates familiar DOM features like hierarchy, events, styling, and querying, which helps streamline the development process. It achieves high-speed rendering across multiple contexts by utilizing concepts such as hoisted scene buffers and virtual DOM trees (SVG). This makes Ripl an efficient tool for creating dynamic visual content.

OnePoseviaGen is an innovative AI tool available as a Hugging Face Space that allows users to upload a single RGB image of an object and instantly generate a 3D representation of it. Beyond just the 3D model, it also estimates the object's full 6D pose, which includes both its position and orientation in space. This capability is particularly useful for applications requiring precise object understanding from limited visual input. The tool operates without the need for multiple views or additional data, streamlining the process of 3D object reconstruction and pose estimation. It is designed for researchers and developers working in computer vision, robotics, and augmented reality fields.

OnePose is a robust code implementation for "One-Shot Object Pose Estimation without CAD Models," a research project featured at CVPR 2022. This tool allows users to estimate the 3D pose of objects from a single image without the need for pre-existing 3D CAD models, a significant advancement in computer vision. It includes comprehensive training and inference code, along with a pipeline to reproduce evaluation results on the OnePose dataset. Users can capture their own training and test data using the OnePose Cap app (iOS only). The project leverages SuperPoint and SuperGlue for 2D feature detection and matching, and COLMAP for Structure-from-Motion. It also offers an optional web-based 3D visualization tool, Wis3D, for interactive analysis of feature matches and estimated poses.

Cad Recode is an innovative AI tool designed to reverse engineer 3D CAD models from mesh files. Users can upload a 3D mesh, and the application processes it by sampling 256 points. From these points, it generates corresponding Python CAD code, effectively translating the visual mesh into programmatic instructions. This generated code can then be used to reconstruct the CAD model, offering a unique approach to converting existing 3D designs into editable, code-based formats. The tool is hosted on Hugging Face Spaces, providing an accessible platform for this specialized conversion task.

EDGS is a Hugging Face Space by CompVis that offers a simplified approach to 3D Gaussian Splatting. Users can upload a front-facing video or a folder of images of a static scene. The tool then automatically extracts frames, and runs a process to optimize the 3D scene. This tool is designed to improve the efficiency of 3D Gaussian Splatting by eliminating the need for densification, making the process more accessible and streamlined for creating 3D representations from 2D inputs. It provides a practical demonstration of the research outlined in the paper "EDGS: Eliminating Densification for Efficient Convergence of 3DGS."

Face Mesh Workflow is a tool hosted on Hugging Face Spaces that allows users to upload an image, detect faces within it, and generate a 3D mesh. It offers the flexibility to adjust depth sources and customize the generated mesh using various sliders. The primary output is an OBJ file, which can then be downloaded for further use in other 3D modeling or animation software. This tool is particularly useful for those working with facial recognition, 3D modeling, or anyone needing to create 3D representations of faces from 2D images.

Playbook offers a secure, production-ready layer built on top of ComfyUI, specifically designed for AI-native studios. It enables these studios to standardize, scale, and protect their generative media pipelines, ensuring consistency and efficiency. The platform allows users to access ComfyUI from any browser, facilitating work from anywhere on any device. Key features include LoRA training and data management, multimodal controls, and tools tailored for media pipelines, helping studios ship mission-critical media projects in days rather than months. Playbook aims to extend creative agency by providing robust control and creativity within generative media workflows.

MeshAnythingV2 is a Hugging Face Space application designed to transform high-detail 3D meshes into optimized, artist-crafted low-poly versions. Users can upload their 3D models in OBJ format and the tool will process them, making them suitable for various applications like game development or stylized rendering. Key features include the ability to enable marching-cubes preprocessing for cleaner geometry and random sampling with a seed to introduce controlled variations in the output. This flexibility allows users to fine-tune the conversion process to achieve desired aesthetic and performance characteristics for their 3D assets. The tool is hosted on Hugging Face, leveraging its infrastructure for accessibility and ease of use.

The 3D Designer Agent is an interactive web application built with Streamlit, hosted on Hugging Face Spaces. This AI tool specializes in transforming text prompts into 3D models, specifically generating printable STL files. It integrates OpenAI for natural language understanding and OpenSCAD for 3D modeling, automating the design process from a simple text description. Users can engage with various functionalities to create custom 3D designs without needing extensive CAD software knowledge. This makes it an accessible solution for individuals looking to quickly visualize and produce physical objects from textual ideas, streamlining the initial stages of 3D design and prototyping.

3D In Your Browser is an innovative AI tool that leverages your webcam to create interactive 3D experiences directly within your web browser. Users can select from a range of pre-defined 3D models, such as a spaceship, parrot, or a Van Gogh-inspired object, and watch them animate and interact in real-time. The application is designed for ease of use, requiring only a webcam to get started. It offers a unique way to visualize and interact with 3D content without the need for complex software installations, making it accessible for quick creative explorations or demonstrations.

The 3D Web Visualizer provides a real-time, interactive 3D representation of a Reachy Mini robot directly in a web browser. This application allows users to rotate, zoom, and pan the robot model, with its joint angles updating dynamically to reflect the robot's actual pose. It's designed for visualizing live robot data, making it suitable for remote monitoring, educational demonstrations, and development purposes. The tool offers a clear and intuitive way to observe robot movements and configurations without needing specialized software, enhancing accessibility for various users.

Align3R is an AI tool available as a Hugging Face Space, designed for generating 3D models from multiple input images. It leverages aligned monocular depth estimation to reconstruct a 3D scene, making it particularly useful for dynamic videos. The tool's core functionality involves analyzing images to estimate depth, which is then used to build a comprehensive 3D representation. This technology is valuable for researchers and developers in computer vision and 3D modeling, offering a practical solution for creating 3D assets from standard image inputs. Its accessibility via Hugging Face Spaces makes it easy to experiment with 3D reconstruction without extensive setup.

Autoforge is an innovative AI tool designed to transform 2D images into 3D printable layered models. Users can upload an image and define their desired filament colors, either manually or by importing CSV or JSON data. The platform offers adjustable parameters to fine-tune the model generation process, ensuring the output meets specific requirements. After processing, users can preview the generated 3D model before downloading a complete zip file containing all necessary files for 3D printing. This tool is particularly useful for designers, artists, and 3D printing enthusiasts looking to convert digital images into tangible, multi-layered physical objects.