AI image generation

This product is a cartoon abstract flat illustration model based on LoRA technology, developed by Beijing Singularity Star Technology Co., Ltd. It focuses on generating cute cartoon-style flat illustrations and is suitable for designers and artists who need to quickly generate illustration materials. Product background information shows that it supports online generation and downloading, and has high user interactivity and community activity. In terms of price, the product provides free trials and paid options, and the specific price is not clearly marked on the page.

GStory is an online video and picture editing platform that provides multiple smart editing features such as background replacement, enhancer, watermark removal and AI image generator. It uses AI technology to simplify the commercial video editing process, improve efficiency, reduce costs, and is trusted by more than 50,000 companies of all sizes.

Project Concept is an AI-based creative conceptualization and mood board tool launched by Adobe, designed to help creative professionals quickly explore and iterate creative concepts in the early stages of a project. It leverages Adobe’s Firefly generative AI model, allowing users to blend images, transform asset regions, blend styles and backgrounds, and more. The product supports multi-person collaboration and integrates with Adobe Creative Cloud, making workflows with Photoshop, Illustrator or Adobe Express seamless. Project Concept also respects the work of artists through Content Credentials technology, ensuring transparency of image provenance and compliant use of AI.

PicLooks is a website that provides free AI-generated real-style avatars, aiming to provide designers and marketers with high-quality avatar images for use in design prototypes, marketing materials, etc. These avatar images are completely generated by AI, avoiding copyright issues that may arise from using images from traditional stock photo websites. Users can quickly find avatars that meet their needs, saving a lot of search time, and can be used for commercial purposes without worrying about copyright issues.

Yaelokre OC Maker is an online platform that allows users to easily create and personalize their original characters. The platform emphasizes customization, user-friendly design and community participation, allowing users to create distinctive characters that reflect their personal style. It offers features such as realistic visuals, user-friendly interface, character diversity, community participation, and 24/7 customer support. Additionally, it provides templates to get started quickly while retaining the flexibility to customize every detail.

Easy Sticker Maker is an artificial intelligence-based online sticker generator that utilizes technologies such as deep learning and generative adversarial networks to generate visually appealing customized stickers based on users' text descriptions. The product supports multiple languages, is easy to use, requires no professional skills, and is suitable for personal and commercial projects. It offers a free trial and has various pricing plans for users to choose from.

Hot Tattoo AI is a revolutionary AI tattoo generator that allows users to easily create custom tattoo designs. Whether looking for the perfect tattoo design for a man or woman, the platform inspires unique and personalized artwork that meets your personal style and preferences. Key advantages of the technology include its intuitive ease of use, rich design variety, ability to quickly generate tattoo designs, and support for collaborative creation with tattoo artists. Additionally, it provides insights into current trends, helping users and artists stay up-to-date.

3D Mesh Generation is an online 3D model generation tool launched by Anything World. It uses artificial intelligence technology to allow users to quickly generate 3D models through simple text descriptions or uploading pictures. The importance of this technology is that it greatly simplifies the 3D model creation process, allowing users without professional 3D modeling skills to easily create high-quality 3D content. Product background information shows that Anything World is committed to providing innovative 3D content creation solutions through its platform, and 3D Mesh Generation is an important part of its product line. Regarding price, users can view specific pricing plans after registration.

AI Comic Factory is an innovative online platform designed to help users easily create unique comics. Users can generate comics in a variety of styles, including options for characters, scenes and dialogue, by entering descriptive prompts or uploading images. The platform offers customization options such as typography layout, character design, and dialogue generation to enhance the visuals of your comics. With user-friendly steps and adjustable settings, achieving high-quality results is easy. Whether for personal entertainment or creative projects, AI Comic Factory makes the comic creation process efficient and enjoyable.

AI Poster Maker AI Poster Maker is an online tool that uses artificial intelligence technology to help users create eye-catching posters without the need for design skills. It automatically generates poster designs through user-described content and selected parameters, greatly simplifying the design process and improving design efficiency. The product background is to meet the market's demand for fast and convenient design tools, and is especially suitable for individuals and companies who need to quickly produce design content. Currently, the product provides a free trial. The specific price and positioning require users to log in to view detailed information.

OC Maker is an online platform that allows users to generate unique and original characters by describing their appearance, personality and special abilities. This AI-driven tool combines creativity and technological innovation to enable users to quickly turn imaginary characters into visual images. Its importance lies in providing a simple and easy-to-use interface, allowing users without professional design skills to create personalized characters, meeting the needs of creative expression and character design. The background information of OC Maker shows that it was developed by a group of technicians who love character design and is designed to help users unleash their creativity and explore different character universes. The product offers a free trial and has different levels of subscription plans to meet the needs of different users.

Free AI Tattoo Generator is an online platform that uses artificial intelligence technology to transform users' text descriptions into unique tattoo designs. It has 29 different tattoo styles for users to choose from and can generate personalized tattoo designs in a short time. The main advantages of this product include unlimited creativity, personalization, diverse style options, speed and efficiency, and cost-effectiveness. Not only does it provide convenience for professional tattoo artists, but it also provides tattoo enthusiasts without drawing skills the opportunity to create professional-quality designs.

Blind Box Studio is a website-based design tool that provides users with an environment to freely explore and build creative workflows through the ComfyUI Workflow platform. This tool supports users to build and modify the design process through node diagrams, and is suitable for a variety of design fields, such as image editing, 3D modeling, etc. The main advantage of Blind Box Studio is its high degree of customization and flexibility, allowing users to create unique workflows based on their needs. Additionally, it supports integration with a variety of plug-ins and extensions, further enhancing its functionality.

Stager AI is a virtual home staging and image editor designed for the real estate industry. It uses artificial intelligence technology to help users quickly and easily enhance the attractiveness of property images, thereby improving the property's online display effect and sales conversion rate. It supports one-click virtual layout, picture enhancement, floor replacement, wall painting, lawn replacement, sky replacement and other functions without requiring users to have professional picture editing skills. Stager AI is designed to help user groups such as real estate agents, photographers and Airbnb hosts to attract more potential buyers or renters by improving the professionalism of property images.

IconGen.io is an AI-powered icon generator that allows users to quickly create beautiful icons suitable for business, presentations, and more. This tool uses artificial intelligence technology to simplify the icon design process, allowing even users without a design background to easily create professional icons. The main advantages of IconGen.io include fast generation, ease of use, and diverse designs, making it suitable for users who need to enhance their brand image.

LOGO123 is a platform that provides professional LOGO design services. It uses artificial intelligence technology to design personalized LOGOs for users, and selects the best design solutions for users through designer PK. The platform supports online submission of design requirements, and users can choose from a variety of package services, including LOGO design, brand VI design, advertising poster design, and trademark registration. LOGO123 is committed to providing users with high-quality, all-round brand design services.

3DTopia-XL is a high-quality 3D asset generation technology built on the Diffusion Transformer (DiT), using a novel 3D representation method PrimX. The technology is capable of encoding 3D shapes, textures and materials into a compact N x D tensor. Each marker is a volumetric primitive anchored on the shape surface, encoding signed distance fields (SDF), RGB and materials with voxelized payloads. This process takes only 5 seconds to generate 3D PBR assets from text/image input, suitable for graphics pipelines.

Adobe Express QR Code Generator is an online tool that allows users to quickly generate scannable QR codes without downloading any software. Users can customize the color and style of QR codes to match their brand or personal style. This tool supports downloading in multiple file formats and is suitable for various scenarios such as business marketing and personal brand promotion. Adobe Express offers a wealth of templates and design resources that make it easy for even novice designers to create eye-catching QR codes.

Phidias is an innovative generative model that utilizes diffusion technology for reference-enhanced 3D generation. The model generates high-quality 3D assets from images, text or 3D conditions and can be completed in seconds. It significantly improves the generation quality, generalization ability and controllability by integrating three key components: Meta-ControlNet that dynamically adjusts the strength of conditions, dynamic reference routing, and self-reference enhancement. Phidias provides a unified framework for 3D generation using text, images and 3D conditions, and has a variety of application scenarios.

Chuzhan is a platform for illustrators, designers and art lovers to display their works, communicate and learn. It brings together the works of many P station (pixiv) painters and certified painters, including anime pictures, anime wallpapers, illustrations, CG original paintings, etc. Touch Station not only provides a space for artists to display their works, but also provides a place for enthusiasts to discover and learn about outstanding works.

Magickimg AI Sticker Generator is an online tool that uses artificial intelligence technology to quickly generate personalized stickers based on prompt words entered by the user. It is mainly aimed at users who need to add a personalized element to social media, chat applications, etc. The product background is based on deep learning technology and provides a simple and fast operating experience through a user-friendly interface. The main advantages of the product include rapid production, easy operation, high-quality output, and safe and reliable service.

Playground: AI Design & Editor is an application that uses artificial intelligence technology to allow users to quickly generate and edit artwork through simple text descriptions. It provides hundreds of pre-designed templates and multiple styles, and users can edit and adjust them unlimited times according to their needs until they achieve a satisfactory design effect. This application is suitable for users who need to quickly design various graphic materials. Whether it is for commercial purposes or personal hobbies, they can easily realize their creativity through Playground's AI tools.

FLUX_UI design is a tool that uses AI technology to assist users in generating artistic UI design models. It is especially suitable for dark UI design on the B-side, supports luminous effects, and can generate a professional and technical e-commerce management interface. This model provides comprehensive operational decision support for merchants through data visualization and business insights.

FLUX Y2K TYPEFACE is a text/font generation model based on LoRA technology, capable of generating Y2K style text, fonts, logos and badges with high accuracy. Licensed by Black Forest Labs, Inc., this model represents a new advancement in text and font generation technology and is highly creative and practical.

AI Icon Generator is an online design tool that uses artificial intelligence technology to help users quickly generate personalized icons. It analyzes user needs and provides diverse icon design options to meet the needs of different design scenarios. This product occupies a place in the design field with its efficiency, ease of use and innovation, and is especially suitable for professionals and teams who need to quickly produce design materials.

DressCode is a text-driven 3D clothing generation framework that aims to democratize design for novices, offering huge potential for fashion design, virtual try-on, and digital human creation. It first introduces SewingGPT, a GPT-based architecture that integrates cross-attention and text-conditional embeddings to generate sewing patterns with text guidance. It then customized a pre-trained Stable Diffusion to generate tile-based Physically Based Rendering (PBR) textures. By leveraging large language models, the framework generates CG-friendly garments through natural language interaction and also facilitates pattern completion and texture editing, streamlining the design process through user-friendly interactions.

Canva Drawable Magic Studio is an online design platform integrated with AI tools, providing a series of design functions from copywriting generation, image editing to animation production. It simplifies the design process, allowing individual creators and teams to quickly realize creative ideas and improve work efficiency. Magic Studio's AI functions include automatic copywriting generation, intelligent image editing, etc., aiming to help users complete high-quality design works in less time and cost.

SceneTeller is an innovative 3D scene generation platform that leverages the latest generative artificial intelligence technology to allow users to design indoor 3D scenes through natural language description. This technology has greatly lowered the technical threshold of 3D design, allowing non-professional users to easily create personalized 3D spaces. Key product benefits include ease of use, high efficiency and creative freedom.

AI Drawing Pro is an application that uses artificial intelligence technology to transform users' text descriptions into visual works of art. It supports a variety of art styles, such as 3D cartoons, animation, oil paintings, watercolors, sketches, Chinese paintings and flat illustrations, etc., providing users with rich creative space. Users can easily create stunning works of art without requiring professional skills, and can easily save and share their creations.

Evined Draw is an innovative drawing application that combines AI technology to bring users a unique drawing experience. Users can draw based on a new theme every day and have unique images generated by the AI ​​model while receiving a rating from 1 to 5 to reflect how well their work fits the theme. This app is suitable for artists of all levels and provides a perfect platform to express their artistic talents.

Adobe Firefly Vector AI is a series of creative generative AI models launched by Adobe, designed to enhance creative work through generative AI capabilities. Firefly models and services are used in Adobe creative applications such as Photoshop, Illustrator, and Lightroom. It helps users generate rich, realistic images and artwork with unprecedented control and creativity through text-to-image, generate fills, generate extensions, and more. Firefly's training data includes Adobe Stock authorized content, publicly licensed content and public domain content to ensure its safety for commercial use. Adobe is committed to developing generative AI responsibly and working closely with the creative community to continuously improve the technology to support and enhance the creative process.

Microsoft Designer is an application developed by Microsoft that uses artificial intelligence technology to help users with creative design and image editing. The application allows users to generate pictures through text descriptions, create personalized wallpapers, design holiday and birthday cards, etc. Its main advantages include powerful AI generation capabilities, ease of use, and rich design templates.

The Social Media Image Generator is an online design tool launched by uBrand Brand Creative Studio. It can help users quickly generate image covers suitable for social media. This tool uses artificial intelligence technology to simplify the design process and improve design efficiency, allowing even novice designers to easily create professional-quality pictures.

AI Logo Designs Gallery is an online platform that uses artificial intelligence technology to provide users with personalized logo design services. Users only need to enter the brand name and some basic design requirements, and AI can generate a variety of logo styles to choose from. The platform supports a variety of industries and styles, including minimalist, medium-complexity, etc., to meet the needs of different users.

Logo Galleria is a platform that uses artificial intelligence technology to provide online logo design. Users simply enter their business name, industry and design preferences to quickly generate a customized logo concept. It provides startups, content creators, and more with an affordable and easy-to-use solution to help them build a strong brand identity.

color4bg.js is a JavaScript library that uses WebGL and JavaScript to generate dynamic, abstract, and visually stunning background images. It allows users to customize up to six colors to generate background patterns, supports dynamic animation effects, and can ensure the same pattern is generated every time through seed values, making it easy to integrate into web design.

ClotheDreamer is a 3D Gaussian-based text-guided clothing generation model that can generate high-fidelity, wearable 3D clothing assets from text descriptions. It uses a novel representation method, Disentangled Clothe Gaussian Splatting (DCGS), which allows clothing and human bodies to be optimized separately. This technology improves the quality of clothing and human body rendering through two-way Score Distillation Sampling (SDS) and supports custom clothing template input. ClotheDreamer's synthetic 3D garments can be easily applied to virtual try-ons and support physically accurate animations.

GENTYPE is an innovative online tool that allows users to convert any image or object into a unique alphabet. This technology not only provides unlimited creative space for designers, but can also be used in education and entertainment to help people learn and express information in novel ways.

MeshAnything is a model that utilizes autoregressive transformers for artist-grade mesh generation that can convert any 3D representation of an asset into artist-created meshes (AMs) that can be seamlessly applied to the 3D industry. It generates meshes with a lower face count, significantly improving storage, rendering, and simulation efficiency while achieving comparable accuracy to previous methods.

Illustration Generator is an AI image generator launched by Icons8, built by a team of professional artists and engineers. It can generate illustrations with a consistent artistic style based on text prompts or reference images provided by users to meet the design needs of web pages, desktop and mobile applications, social media, marketing, etc. The AI ​​model is trained based on visual materials produced by Icons8 artists, ensuring that the generated AI artwork has a consistent and professional appearance.

FontStudio is an innovative font effect generation model that uses adaptive diffusion technology to generate coherent visual content on irregular font shape canvases. This technology breaks through the limitations of traditional rectangular canvases and provides new solutions for multi-language font design. The FontStudio system has shown clear advantages in user preference studies, even achieving a 78% aesthetic win rate when compared to top commercial products such as Adobe Firefly.

SketchDeco is an innovative online tool that converts black and white sketches, masks and color palettes into realistic color images without the need for user-defined text prompts. This technology combines ControlNet and a staged generation approach to provide faithful image generation and user-directed colorization using Stable Diffusion v1.5 and BLIP-2 text hints. Not only is it fast and requires no training, it's also compatible with consumer-grade Nvidia RTX 4090 Super GPUs, providing a valuable resource for creative professionals and enthusiasts.

MaPa is an innovative approach to generating materials for 3D meshes based on textual descriptions. This technology creates segmented procedural material maps to represent appearance, enables high-quality rendering, and provides significant flexibility in editing. Utilizing pre-trained 2D diffusion models, MaPa bridges the gap between textual descriptions and material maps without requiring large amounts of paired data. This technology decomposes the shape into multiple parts and designs a diffusion model of the control segments to synthesize a 2D image aligned with the mesh part, then initializes the parameters of the material map and fine-tunes it through the differentiable rendering module to produce a material that conforms to the text description. Extensive experiments show that MaPa outperforms existing technologies in terms of fidelity, resolution, and editability.

Sticker Creator is an online tool launched by Microsoft that allows users to create personalized stickers and images. It supports a variety of styles and elements, such as 3D rendering, pixel art, cartoon style, etc., and users can generate unique visual content according to their needs. The tool comes amid Microsoft's commitment to delivering innovative design solutions that meet users' needs in digital media and creative expression.

TimeUi is a custom timeline node system designed for ComfyUI, aiming to create timelines similar to video/animation editing tools, but without relying on traditional timecode. Users can easily add, delete or rearrange rows, providing a smooth user experience. The system supports image upload and management, allowing users to upload images directly to nodes or attach other "upload image" nodes, simplifying the workflow. In addition, each timeline row includes a variety of customization settings, such as toggling the visibility of image masks and increasing control over image adjustments. Nodes can work independently or with other external nodes, easily switch settings such as IP adapters, image negatives, attention masks, clip vision, masks and more to fine-tune the output.

DreamMat is an innovative model that generates Physically Based Rendering (PBR) materials for 3D meshes based on text prompts. It solves the shortcomings of existing 2D diffusion models in material decomposition and generates high-quality PBR materials that are consistent with the given geometry and lighting environment and have no built-in shadow effects. This technology is of great significance for downstream tasks such as gaming and film production, as it significantly improves rendering quality and enhances the user's visual experience.

Stylar AI's 2D to 3D Image Converter is a powerful image conversion tool that utilizes advanced Image-to-Image technology to convert flat 2D images into 3D images. This tool provides high-quality image conversion and multiple style options to meet users' needs for 3D rendering of images. The main functions of the product include uploading pictures, selecting 3D effects, downloading 3D creations, etc. It also provides a variety of 3D styles, such as 3D cartoon effects, 3D artwork, etc., as well as the function of converting sketches into 3D designs.

Meijian AI Creative Commercial Photography is an online design platform focusing on the field of e-commerce. It uses artificial intelligence technology to help users quickly generate various design drawings required for e-commerce, such as main images, header images, etc. The core advantage of this platform lies in its efficient design generation speed and rich template resources, which can meet the design needs of different categories of goods, thereby improving the efficiency and effectiveness of e-commerce operations.

Boords is an AI character generator that simplifies the character design process by helping users create consistent, identifiable AI characters and place them in any scene. It provides powerful character editing functions, and users can easily customize the appearance and characteristics of the character. Boords also supports the rapid generation of character variants with different clothing and expressions to meet various scene needs. Users can create interesting storyboards using simple image sequences and easily build compelling character-driven stories. Boords is trusted by 650,000 professionals on the world's top video teams to dramatically simplify their pre-production processes.

Interactive3D is an advanced 3D generative model that provides users with precise control through interactive design. The model adopts a two-stage cascade structure, utilizing different 3D representation methods, allowing the user to modify and guide at any intermediate step of the generation process. Its importance lies in enabling users to have fine control over the 3D model generation process, thereby creating high-quality 3D models that meet specific needs.

The intelligent comic creation tool launched by Baidu Wenku has the following advantages: 1. One-stop creation process, seamless connection from idea to finished product. 2. A variety of comic styles are available, such as Korean cartoons, romantic thick paint, etc. 3. The characters, expressions, scenes, etc. are described with fine detail and vividness, comparable to the level of cartoonists. 4. Built-in intelligent editor can refine details. 5. No professional painting skills are required, you can create with creativity. 6. It is currently in the public beta stage, you can follow the official account to make an appointment to experience it.

The StableDesign project aims to provide data sets and training methods for generative interior design. Users upload pictures of empty rooms and text prompts to generate decoration renderings. Through Airbnb data download, feature extraction and ControlNet model training, combined with image processing and natural language processing technology, new ideas and methods are provided.

Adobe Express GPT is a customized GPT provided by Adobe for ChatGPT Plus users. It can quickly match more than 200,000 professional design templates of Adobe Express according to the user's text prompts, and import the matching results directly into Adobe Express for further editing and publishing. The plug-in also integrates generative AI functions such as Adobe Firefly, such as text to image, object removal, etc., helping users easily create outstanding visual content. Adobe is committed to developing AI technology responsibly, and Firefly models are trained using authorized content and can be safely used for commercial purposes.

GRM is a large-scale reconstruction model that can recover 3D assets from sparse view images in 0.1 seconds and generate them in 8 seconds. It is a feed-forward Transformer-based model that can efficiently fuse multi-view information to convert input pixels into pixel-aligned Gaussian distributions. These Gaussian distributions can be back-projected into a dense 3D Gaussian distribution collection representing the scene. Our Transformer architecture and use of 3D Gaussian distribution unlocks a scalable and efficient reconstruction framework. Extensive experimental results demonstrate the superiority of our approach over other alternatives in terms of reconstruction quality and efficiency. We also demonstrate the potential of GRM in generative tasks such as text to 3D and image to 3D, by combining with existing multi-view diffusion models.

Stable Video 3D is a new model launched by Stability AI, which has made significant progress in the field of 3D technology, providing greatly improved quality and multi-view support compared to the previously released Stable Zero123. The model is able to generate orbital videos based on a single image input without camera conditions, and is able to create 3D videos along specified camera paths.

DragAnything is a product that uses entity representation to achieve motion control of arbitrary objects. Unlike the traditional way of dragging pixels, DragAnything enables true entity-level motion control. It can realize user trajectory interaction and has SAM function. The product can precisely control the movement of objects and generate high-quality videos, and users only need to draw a trajectory during the interaction. DragAnything enables diverse motion control of different elements such as foreground, background and camera. Positioned in the design field, it is suitable for scenes that require fine control of objects in videos. Product pricing was not disclosed.

3D AI Studio is an online tool based on artificial intelligence technology that can easily generate customized 3D models. Suitable for designers, developers and creative people, providing high-quality digital assets. Users can quickly create 3D models through the AI ​​generator and export them in FBX, GLB or USDZ format. 3D AI Studio features high performance, user-friendly interface, and automatic generation of real textures, which can significantly shorten modeling time and reduce costs.

Leonardo.Ai Realtime Canvas is a real-time intelligent drawing tool. It uses AI technology to instantly convert simple sketches into high-quality images, greatly improving designers' creative efficiency. Compared with the traditional design process, Leonardo.Ai allows designers to skip the tedious steps of manual drawing and post-processing and focus on the creativity itself. Key functions include: real-time drawing conversion, image enhancement, intelligent expansion, etc. Suitable for graphic design, illustration design, UI design and other fields.

ComfyUI-Mana-Nodes is a set of custom nodes designed for ComfyUI, including the ability to convert fonts into image animations. Users can create dynamic images and video effects through these nodes. The project follows the MIT license and encourages community contributions and customization.

ComfyUI-3D-Pack is a powerful collection of 3D processing plug-ins. It provides ComfyUI with the ability to process 3D models (grids, textures, etc.), and integrates various cutting-edge 3D reconstruction and rendering algorithms, such as 3D Gaussian sampling, NeRF different iable rendering, etc., which can quickly reconstruct 3D Gaussian models from single-view images and convert them into triangular mesh models. It also provides an interactive 3D visualization interface.

Keyframer is a prototype animation generation tool based on a large language model developed by Apple. It can automatically add animation effects to SVG images through text descriptions and convert them into CSS codes. Users without programming experience can simply upload images and enter text descriptions, and Keyframer will automatically generate code. Compared with other AI-generated animation solutions, Keyframer is simpler and easier to use. It's still in the prototype stage and public availability remains to be seen.

ComfyUI-3D-Pack is a powerful 3D processing node plug-in package. It provides ComfyUI with the ability to process 3D inputs (grids, UV textures, etc.), using the most cutting-edge algorithms, such as 3D Gaussian sampling, neural radiation fields, etc. This project allows users to quickly generate a 3D Gaussian model using only a single image, and convert the Gaussian model into a grid to achieve 3D reconstruction. It also supports multi-view images as input, allowing texture maps for multi-view rendering to be mapped on a given 3D mesh. The plug-in package is under development and has not yet been officially released to the ComfyUI plug-in library, but it already supports functions such as large multi-view Gaussian models, three-plane Gaussian transformers, 3D Gaussian sampling, depth mesh triangulation, 3D file loading and saving, etc. It aims to be a powerful tool for ComfyUI to handle 3D content.

Glif StyleHunter is a Chrome browser extension that allows you to select any image on a web page and generate mixed images in various styles based on your prompt words. Simply right-click on an image and enter your prompt word to apply that image style directly to a new image you want to create, whether to imitate that style or combine it with other styles to create something entirely new. This extension provides users with an intuitive and flexible way to explore and experiment with different visual art styles, as well as apply these styles to their own creative projects.

BlockFusion is a diffusion-based model that generates 3D scenes and seamlessly integrates new blocks into the scene. It is trained on a dataset of 3D patches randomly cropped from a complete 3D scene mesh. Through block-by-block fitting, all training blocks are converted into hybrid neural fields: triahedrons containing geometric features, followed by a multilayer perceptron (MLP) for decoding signed distance values. A variational autoencoder is used to compress the triahedrons into a latent trihedral space, subjecting them to a denoising diffusion process. Diffusion is applied to latent representations, which enables high-quality and diverse 3D scene generation. When extending a scene during generation, simply append empty blocks to overlap the current scene and extrapolate existing potential triahedrons to fill the new blocks. Extrapolation is accomplished by tuning the generation process using feature samples from overlapping triahedrons during denoising iterations. Latent trihedral extrapolation produces semantically and geometrically meaningful transitions that blend harmoniously with the existing scene. Use the 2D layout adjustment mechanism to control the placement and arrangement of scene elements. Experimental results show that BlockFusion is capable of generating diverse, geometrically consistent, and high-quality indoor and outdoor large-scale 3D scenes.

StrokeNUWA is a pioneering work that explores better visual representation of "partition markup" on vector graphics that is visually semantically rich, naturally compatible with LLMs, and highly compressible. Equipped with partition markers, StrokeNUWA significantly outperforms traditional LLM-based and optimization-based methods on various metrics for vector graphics generation tasks. In addition, StrokeNUWA achieves up to 94x acceleration in inference speed and has an excellent SVG code compression ratio of 6.9% compared to previous methods.

Media2Face is a co-linguistic facial animation generation tool guided by audio, text and image multimodality. It first utilizes General Neural Parameterized Facial Assets (GNPFA) to map facial geometry and images into a highly general expression latent space, and then extracts high-quality expressions and accurate head poses from a large number of videos to construct the M2F-D dataset. Finally, the diffusion model in the GNPFA latent space is adopted for co-linguistic facial animation generation. The tool not only delivers high fidelity in facial animation synthesis, but also expands expressiveness and style adaptability.

InternLM-XComposer2 is a leading visual language model, good at free-form text image synthesis and understanding. The model not only understands traditional visual language, but also expertly constructs intertwined text-image content from a variety of inputs, such as outlines, detailed text specifications, and reference images, enabling highly customizable content creation. InternLM-XComposer2 proposes a Partial LoRA (PLoRA) approach that specifically applies additional LoRA parameters to image tagging to preserve the integrity of pre-trained language knowledge and achieve a balance between precise visual understanding and literary composition of texts. Experimental results show that InternLM-XComposer2 based on InternLM2-7B is superior in generating high-quality long text multi-modal content, as well as its excellent visual language understanding performance in various benchmark tests, not only significantly better than existing multi-modal models, but also on par with or even exceeding GPT-4V and Gemini Pro in some evaluations. This highlights its excellent capabilities in the field of multi-modal understanding. The InternLM-XComposer2 series of models has 7B parameters and is publicly available at https://github.com/InternLM/InternLM-XComposer.

Comfy Textures is an Unreal Engine plug-in that integrates the editor with ComfyUI, allowing you to quickly create and adjust the texture of your scene using a generative diffusion model. Supports single-viewpoint and multi-viewpoint texture projection, and can be used with perspective and orthographic cameras. Texture editing and image-to-image workflows are also supported. Works seamlessly with Unreal Engine 5.x and 4.x.

CreativeSynth is an innovative unified framework based on the diffusion model with the ability to coordinate multi-modal input and multi-tasking. By combining multi-modal features with a customized attention mechanism, CreativeSynth enables the introduction of realistic semantic content into the artistic field, precisely manipulating image style and content through inversion and real-time style transfer, while maintaining the integrity of the original model parameters. Rigorous qualitative and quantitative evaluation highlights CreativeSynth's strengths in enhancing the fidelity of artistic images while preserving their inherent aesthetic essence. By bridging the gap between generative models and artistic essence, CreativeSynth becomes a custom digital palette.

3DTopia is a two-stage text-to-3D generative model. The first stage uses a diffusion model to quickly generate candidates. The second stage optimizes the assets selected in the first stage. This model enables high-quality text-to-3D generation in under 5 minutes.

Davinci Pencil is a drawing application tailor-made for iPad. With the help of artificial intelligence, you can unlock the potential of your drawing skills and turn your wildest ideas and fantasies into beautiful paintings. Our easy-to-use interface streamlines your drawing experience, allowing you to imagine, draw, and doodle in seconds. We update our database frequently to constantly bring you fresh aesthetics. The main functions include: AI rendering function, detail configuration, drawing overlay on rendered images, drawing comparison, viewing all drawing images, creating multiple pages, and selecting custom sizes. Enhance your drawings with Davinci Pencil and turn your imagination into paintings!

Make-A-Shape is a new 3D generative model designed to train on large-scale data in an efficient manner, capable of leveraging 10 million publicly available shapes. We innovatively introduce a wavelet tree representation to compactly encode the shape by formulating a subband coefficient filtering scheme, and then arrange the representation in a low-resolution grid by designing a subband coefficient packing scheme, making it generative of diffusion models. Furthermore, we propose a subband adaptive training strategy that enables our model to effectively learn to generate coarse and fine wavelet coefficients. Finally, we extend our framework to be controlled by additional input conditions to enable it to generate shapes from various modalities, such as single/multi-view images, point clouds, and low-resolution voxels. In extensive experiments, we demonstrate various applications such as unconditional generation, shape completion, and conditional generation. Our method not only surpasses the state of the art in providing high-quality results, but also efficiently generates shapes in seconds, typically only 2 seconds under most conditions.

RPG-DiffusionMaster is a new training-free text-to-image generation/editing framework that utilizes the chain reasoning capabilities of multi-modal LLM to enhance the combinability of text-to-image diffusion models. This framework uses MLLM as a global planner to decompose the complex image generation process into simple generation tasks within multiple sub-regions. At the same time, complementary regional diffusion is proposed to achieve regionalized combination generation. Furthermore, text-guided image generation and editing are integrated in a closed-loop manner in the proposed RPG framework, thereby enhancing the generalization ability. Extensive experiments prove that RPG-DiffusionMaster outperforms state-of-the-art text-to-image diffusion models such as DALL-E 3 and SDXL in terms of multi-category object combination and text-image semantic alignment. In particular, the RPG framework is broadly compatible with various MLLM architectures (e.g. MiniGPT-4) and diffusion backbones (e.g. ControlNet).

AnimatableDreamer is a framework for generating and reconstructing animatable non-rigid 3D models from monocular videos. It is able to generate different categories of non-rigid objects while following object motions extracted from videos. The key technology is the proposed canonical fraction distillation method, which simplifies the generation dimension from 4D to 3D, performs noise reduction on different frames in the video, and performs the distillation process in a unique canonical space. This ensures temporally consistent generation and morphological fidelity in different poses. With the help of differentiable deformation, AnimatableDreamer upgrades the 3D generator to 4D, providing a new perspective for the generation and reconstruction of non-rigid 3D models. Furthermore, combined with the inductive knowledge of the consistency diffusion model, canonical fractional distillation can regularize the reconstruction from a new perspective, thereby closing the loop and enhancing the generation process. Extensive experiments show that this method can generate highly flexible text-guided 3D models from monocular videos, while the reconstruction performance is better than typical non-rigid body reconstruction methods.

HexaGen3D is an innovative method for generating high-quality 3D assets from text prompts. It leverages a large pre-trained 2D diffusion model by fine-tuning a pre-trained text-to-image model to jointly predict 6 orthogonal projections and corresponding latent triahedrons, and then decodes these latent values ​​to generate texture meshes. HexaGen3D does not require per-sample optimization and can infer high-quality and diverse objects from text prompts in 7 seconds, providing a better quality and latency trade-off compared to existing methods. In addition, HexaGen3D has strong generalization capabilities to new objects or combinations.

InseRF is a method for generating new objects in NeRF-reconstructed 3D scenes via text cues and 2D bounding boxes. It generates new 3D objects from a user-supplied text description and a 2D bounding box in a reference viewpoint and inserts them into the scene. This method enables controlled, 3D-consistent object insertion without the need for explicit 3D information. Through experiments in multiple 3D scenes, the effectiveness of the InseRF method relative to existing methods is demonstrated.

URHand is the first universal illuminated hand model that generalizes across different views, poses, lighting and identities. The model can be personalized with fewer shots using images taken with the phone, and can be realistically rendered in new lighting conditions. We construct a powerful universal lighting prior based on illumination of multi-view hand images using neural networks. We propose a neural renderer that employs a spatially varying linear lighting model with physically inspired shadows as input features. Our specifically designed lighting model explicitly maintains the linearity of light transmission by removing nonlinear activation and bias. We also introduce joint learning of physically based models and neural lighting models to further improve fidelity and generalization performance. Extensive experiments show that our method outperforms existing methods in terms of quality and generalization ability. We also show how to quickly personalize unseen identities from mobile phones.

Make-A-Character (Mach) is a user-friendly framework designed to create lifelike 3D avatars from text descriptions. The framework leverages the power of large language and visual models for text intent understanding and intermediate image generation, and then goes through a series of human-oriented visual perception and 3D generation modules. Our system provides an intuitive way for users to create controllable, realistic, fully realized 3D characters in 2 minutes, while also easily integrating with existing CG pipelines to achieve dynamic performance.

This product is a 3D GAN technology that can parse fine-grained 3D geometry with unprecedented detail by learning a method based on neural volume rendering. The product uses a learning sampler to accelerate 3D GAN training, use less depth sampling, and directly render each pixel of the full-resolution image during training and inference. At the same time, it learns high-quality surface geometry and synthesizes images with high-resolution 3D geometry and strict viewing angles. The product demonstrates state-of-the-art 3D geometry quality at FFHQ and AFHQ, setting a new standard for unsupervised learning in 3D GANs.

Steerable Motion is a ComfyUI node for batch creative interpolation. Our goal is to show how to best drive motion as the image evolves as a video model.

Art QR Code is a tool that uses AI painting technology to generate artistic style QR codes. Users can upload their own QR code pictures and choose different artistic styles to generate unique artistic QR codes. Artistic QR codes can be used in personal blogs, social media, business cards and other scenarios to help users enhance their brand image.

Jinshu is an innovative artistic word generation tool that provides rich artistic font styles and effects. Users can quickly generate personalized artistic word works. This tool is positioned to provide users with convenient and efficient word art generation services, and can produce exquisite word art works without professional design skills.

SceneWiz3D is a novel method for synthesizing high-fidelity 3D scenes from text. It uses a hybrid 3D representation, with explicit representation for objects and implicit representation for scenes. Users can generate objects through traditional text-to-3D methods or by providing objects themselves. To configure the scene layout and automatically place objects, we applied particle swarm optimization technology during the optimization process. Furthermore, in the text-to-scene case, it is difficult to obtain multi-view supervision for certain parts of the scene (e.g., corners, occlusions), resulting in inferior geometry. To alleviate this lack of supervision, we introduce the RGBD panoramic diffusion model as an additional prior, thus achieving high-quality geometries. Extensive evaluation supports that our method achieves higher quality than previous methods, generating detailed and perspective-consistent 3D scenes.

roomGPT is an online service that allows you to upload room photos and use AI technology to generate ideal room effects. Users only need to upload photos of existing rooms, and the system can generate room design effects of different styles for users to choose their favorite style. The service uses the ControlNet machine learning model to generate different variations of the room. The free open source version of roomGPT can be deployed locally and also provides paid SaaS services.

Stable Zero123 is an in-house trained model for view-conditioned image generation. Stable Zero123 produces significantly improved results compared to its previous cutting-edge technology, Zero123-XL. It achieves this goal through three key innovations: 1. An improved training dataset that is heavily filtered from the Objaverse, retaining only high-quality 3D objects and rendering more realistically than previous methods. 2. During training and inference, we provide the model with estimated camera angles. This elevation condition allows it to make more informed, higher quality predictions. 3. Pre-computed datasets (pre-computed latent variables) and an improved data loader that supports higher batch sizes, coupled with the first innovation, make training 40 times more efficient than Zero123-XL. The model is now available on Hugging Face for researchers and non-commercial users to download and experiment with.

GPT Chart Maker is an artificial intelligence chart generation tool based on ChatGPT that can create impressive charts and graphs in seconds, making it easy to get professional infographics.

A collection of 3D Gaussian splatter technology resources, covering ecosystem and tools, research papers, Unity Gaussian scattering projects, etc. This technology is widely used in 3D editing, real-time point cloud relighting, inverse rendering, data compression, anti-aliasing and other fields. It has a high reference value for people interested in 3D Gaussian splatter technology.

Generative Powers of Ten is an approach that leverages text-to-image models to generate multi-scale consistent content, enabling extreme semantic scaling of scenes, such as from a wide-angle landscape view of a forest to a macro shot of an insect on a tree branch. This representation enables us to render continuously zoomed videos, or to interactively explore different scales of a scene. We achieve this through a joint multiscale diffusion sampling approach that encourages consistency across scales while preserving the integrity of each individual sampling process. Since each generated scale is guided by a different textual hint, our method is able to achieve deeper scaling than traditional super-resolution methods, which may struggle to create new contextual structures at completely different scales. We qualitatively compare our approach on alternative techniques for image super-resolution and external rendering, and show that our approach is most effective at generating consistent multi-scale content.

Spline AI is a tool for quickly generating 3D objects, animations and textures through AI. Using simple tips, designers can turn ideas into reality faster. Product functions include: generating 3D objects and scenes, editing objects, applying materials, adding lighting, generating seamless textures, etc. Spline AI also provides AI texture functionality that can generate seamless textures based on text prompts. This product is suitable for designers, artists and creative teams.

MeshGPT creates triangular meshes by autoregressively sampling from a transformer model trained to generate labels from a learned geometric vocabulary. These markers can then be decoded into faces of a triangular mesh. Our method generates clean, coherent and compact meshes with sharp edges and high fidelity. MeshGPT performs significantly better in shape coverage than existing mesh generation methods, with FID scores improving by 30 points across various categories.

4D-fy is a text-to-4D generation method that uses mixed fractional distillation sampling technology and combines the supervision signals of multiple pre-trained diffusion models to achieve high-fidelity text-to-4D scene generation. Its approach parametrizes 4D radiation fields through neural representations, uses static and dynamic multi-scale hash table features, and utilizes volume rendering to render images and videos from the representations. By mixed fractional distillation sampling, gradients of a 3D-aware text-to-image model (3D-T2I) are first used to optimize representation, then gradients of a text-to-image model (T2I) are combined to improve appearance, and finally gradients of a text-to-video model (T2V) are combined to increase the motion of the scene. 4D-fy can generate 4D scenes with attractive appearance, 3D structure and motion.

3D Paintbrus is a technology that automatically adds texture to local semantic areas on a mesh via text description. This method operates directly on the mesh, producing texture maps that are seamlessly integrated into the standard graphics pipeline. At the same time, a localized map of the specified editing area and a matching texture map are generated. We utilize multiple stages of a cascaded diffusion model to oversee local editing techniques that enhance detail and resolution in textured areas. The technique, called Cascade Fractional Distillation (CSD), is able to simultaneously distill scores at multiple resolutions in a cascade manner, enabling control over the granularity and global understanding of supervision. We demonstrate the effectiveness of 3D brushes in locally texturing various shapes within different semantic regions.

LiveSketch is a tool for adding animation effects to hand-drawn sketches. It can automatically generate vector animations based on text prompts to bring sketches to life. The tool does not require complex training and uses a pre-trained text-to-video model to guide the movement of strokes. It is suitable for designers, animators and other users who need to add animation effects to sketches. Animated drawings can be used on the website.

draw-fast is a fast drawing tool developed using TypeScript, CSS and JavaScript. It provides fast drawing functions and has the advantage of simplicity and ease of use. Positioned to provide users with fast drawing solutions.

ZipLoRA is a method that effectively merges independently trained style and topic LoRAs to achieve content generation under any user-provided topic and style. Through optimized methods, ZipLoRA is able to retain the content and style-generating properties of original LoRAs, while being able to re-contextualize reference objects and have the ability to control the degree of style. This approach achieves significant improvements in thematic and stylistic fidelity.

FlyArt is the ultimate graphic design app for easily creating personalized flyers. We offer a variety of ready-made flyer templates that you can customize according to your needs. With FlyArt, you can choose the right designs and customize them, adding background images, stickers, logos, fonts and brand colors. FlyArt helps your business stand out on social media and provide rapid growth.

WxArt Ai is a professional painting software with a powerful AI engine that provides users with a series of innovative features. Utilizing advanced AI content generation technology, WxArt Ai can create a variety of pictures, including text-based art and image-based art. Whether you are looking for ink style, colorful animation, photorealistic style or 2D work, WxArt Ai can meet your diverse needs.

Innovative WordArt is a tool that can deform and generate textures for text through customized concepts. This tool allows users to build creative, personalized glyphs and textures. The tool has an easy-to-use interface and a variety of font and texture options to meet users' needs for text expression in designs. Innovative artistic words are positioned to provide creative inspiration and design elements for designers, artists, etc.

AI Stickers is a service that turns your text into vivid, personalized stickers. Say goodbye to monotonous messages and express yourself on platforms like WhatsApp, Telegram and more with AI-generated stickers. Unleash your creativity and bring conversation to life.

Genie is a research preview of Luma's 3D generative base model. It can generate various three-dimensional models for use in design, creation, entertainment and other fields. Genie provides a wealth of features, including shape generation, texture drawing, animation creation, and more. It can be used in many fields such as game development, virtual reality, and movie special effects. Genie pricing and positioning will be determined ahead of official launch.