Close Menu
Ztoog
    What's Hot
    Science

    Smart Pillows: Sweet Dreams Are Made of This

    Mobile

    Sharing photos on Google Messages could get a whole lot easier

    Science

    $158,000 ALS drug pulled from market after failing in large clinical trial

    Important Pages:
    • About Us
    • Contact us
    • Privacy Policy
    • Terms & Conditions
    Facebook X (Twitter) Instagram Pinterest
    Facebook X (Twitter) Instagram Pinterest
    Ztoog
    • Home
    • The Future

      Any wall can be turned into a camera to see around corners

      JD Vance and President Trump’s Sons Hype Bitcoin at Las Vegas Conference

      AI may already be shrinking entry-level jobs in tech, new research suggests

      Today’s NYT Strands Hints, Answer and Help for May 26 #449

      LiberNovo Omni: The World’s First Dynamic Ergonomic Chair

    • Technology

      A Replit employee details a critical security flaw in web apps created using AI-powered app builder Lovable that exposes API keys and personal info of app users (Reed Albergotti/Semafor)

      Gemini in Google Drive can now help you skip watching that painfully long Zoom meeting

      Apple iPhone exports from China to the US fall 76% as India output surges

      Today’s NYT Wordle Hints, Answer and Help for May 26, #1437

      5 Skills Kids (and Adults) Need in an AI World – O’Reilly

    • Gadgets

      Future-proof your career by mastering AI skills for just $20

      8 Best Vegan Meal Delivery Services and Kits (2025), Tested and Reviewed

      Google Home is getting deeper Gemini integration and a new widget

      Google Announces AI Ultra Subscription Plan With Premium Features

      Google shows off Android XR-based glasses, announces Warby Parker team-up

    • Mobile

      Deals: the Galaxy S25 series comes with a free tablet, Google Pixels heavily discounted

      Microsoft is done being subtle – this new tool screams “upgrade now”

      Wallpaper Wednesday: Android wallpapers 2025-05-28

      Google can make smart glasses accessible with Warby Parker, Gentle Monster deals

      vivo T4 Ultra specs leak

    • Science

      Analysts Say Trump Trade Wars Would Harm the Entire US Energy Sector, From Oil to Solar

      Do we have free will? Quantum experiments may soon reveal the answer

      Was Planet Nine exiled from the solar system as a baby?

      How farmers can help rescue water-loving birds

      A trip to the farm where loofahs grow on vines

    • AI

      Rationale engineering generates a compact new tool for gene therapy | Ztoog

      The AI Hype Index: College students are hooked on ChatGPT

      Learning how to predict rare kinds of failures | Ztoog

      Anthropic’s new hybrid AI model can work on tasks autonomously for hours at a time

      AI learns how vision and sound are connected, without human intervention | Ztoog

    • Crypto

      GameStop bought $500 million of bitcoin

      CoinW Teams Up with Superteam Europe to Conclude Solana Hackathon and Accelerate Web3 Innovation in Europe

      Ethereum Net Flows Turn Negative As Bulls Push For $3,500

      Bitcoin’s Power Compared To Nuclear Reactor By Brazilian Business Leader

      Senate advances GENIUS Act after cloture vote passes

    Ztoog
    Home » Researchers from ByteDance and CMU Introduce AvatarVerse: A Novel AI Pipeline for Generating High-Quality 3D Avatars Controlled by both Text Descriptions and Pose Guidance
    AI

    Researchers from ByteDance and CMU Introduce AvatarVerse: A Novel AI Pipeline for Generating High-Quality 3D Avatars Controlled by both Text Descriptions and Pose Guidance

    Facebook Twitter Pinterest WhatsApp
    Researchers from ByteDance and CMU Introduce AvatarVerse: A Novel AI Pipeline for Generating High-Quality 3D Avatars Controlled by both Text Descriptions and Pose Guidance
    Share
    Facebook Twitter LinkedIn Pinterest WhatsApp

    3D avatars have in depth use in industries together with sport improvement, social media and communication, augmented and digital actuality, and human-computer interplay. The building of high-quality 3D avatars has attracted a whole lot of curiosity. These complicated 3D fashions are historically constructed manually, which is a labor-intensive and time-consuming process that takes hundreds of hours from educated artists with substantial aesthetic and 3D modeling data. As a consequence, their work’s goal is to automate the creation of high-quality 3D avatars utilizing solely pure language descriptions as a result of this has vital analysis potential and the flexibility to preserve sources. 

    Reconstructing high-fidelity 3D avatars from multi-view movies or reference pictures has garnered a lot consideration just lately. These methods can’t assemble imaginative avatars with difficult textual content prompts since they depend on restrictive visible priors obtained from movies or reference photos. Diffusion fashions show spectacular ingenuity when creating 2D pictures, largely as a result of many large-scale text-image mixtures can be found. However, the shortage of variety and scarcity of 3D fashions make it tough to coach a 3D diffusion mannequin adequately. 

    Recent analysis has regarded into optimizing Neural Radiance Fields for producing high-fidelity 3D fashions utilizing pre-trained text-image generative fashions. However, creating stable 3D avatars with numerous positions, seems, and kinds are nonetheless difficult. For occasion, utilizing widespread rating distillation sampling with out further management to direct NeRF optimization will seemingly introduce the Janus concern. Aside from that, the avatars created by the current strategies ceaselessly show observable coarseness and blurriness, which ends up in the absence of high-resolution native texture particulars, equipment, and different vital points. 

    Researchers from ByteDance and CMU counsel AvatarVerse, a novel framework made for producing high-quality and dependable 3D avatars utilizing textual descriptions and place guidances, to handle these limitations. They initially practice a brand-new ControlNet utilizing 800K or extra human DensePose photos. Then, on prime of the ControlNet, SDS loss conditional on the 2D DensePose sign is carried out. They can obtain precise view correspondence between each 2D view and the 3D area and between many 2D views. Their expertise does away with the Janus drawback that plagues nearly all of earlier approaches whereas additionally enabling pose management of the created avatars. As a consequence, it ensures a extra dependable and constant era process for avatars. The produced avatars can also be effectively aligned with the joints of the SMPL mannequin due to the exact and adaptable supervision indicators supplied by DensePose, making skeletal binding and management simple and environment friendly.

    They current a progressive high-resolution era approach to enhance the realism and element of native geometry, whereas simply counting on DensePose-conditioned ControlNet could produce native artifacts. They use a smoothness loss, which regularises the synthesis course of by selling a smoother gradient of the density voxel grid inside their computationally efficient express Neural Radiance Fields to cut back the coarseness of the created avatar. 

    These are the general contributions: 

    • They introduce AvatarVerse, a way that enables a high-quality 3D avatar to be routinely created utilizing solely a phrase description and a reference human stance. 

    • They present the DensePose-Conditioned Score Distillation Sampling Loss, a way that makes it simpler to create pose-aware 3D avatars and efficiently mitigates the Janus drawback, bettering system stability. 

    • Through a methodical high-resolution producing course of, they enhance the standard of the generated 3D avatars. This expertise creates 3D avatars with distinctive element, together with fingers, equipment, and extra, by means of a rigorous coarse-to-fine refinement course of. 

    • AvatarVerse performs admirably, outperforming rivals in high quality and stability. AvatarVerse’s superiority in creating high-fidelity 3D avatars is demonstrated by meticulous qualitative assessments supported by thorough person analysis. 

    This units a brand new customary for dependable, zero-shot 3D avatar era of the very best caliber. They have put up demos of their approach on their GitHub web site.


    Check out the Paper and GitHub. All Credit For This Research Goes To the Researchers on This Project. Also, don’t neglect to hitch our 28k+ ML SubReddit, 40k+ Facebook Community, Discord Channel, and Email Newsletter, the place we share the most recent AI analysis information, cool AI tasks, and extra.


    Aneesh Tickoo is a consulting intern at MarktechPost. He is at the moment pursuing his undergraduate diploma in Data Science and Artificial Intelligence from the Indian Institute of Technology(IIT), Bhilai. He spends most of his time engaged on tasks geared toward harnessing the facility of machine studying. His analysis curiosity is picture processing and is keen about constructing options round it. He loves to attach with individuals and collaborate on attention-grabbing tasks.


    🔥 Use SQL to foretell the long run (Sponsored)

    Share. Facebook Twitter Pinterest LinkedIn WhatsApp

    Related Posts

    AI

    Rationale engineering generates a compact new tool for gene therapy | Ztoog

    AI

    The AI Hype Index: College students are hooked on ChatGPT

    AI

    Learning how to predict rare kinds of failures | Ztoog

    AI

    Anthropic’s new hybrid AI model can work on tasks autonomously for hours at a time

    AI

    AI learns how vision and sound are connected, without human intervention | Ztoog

    AI

    How AI is introducing errors into courtrooms

    AI

    With AI, researchers predict the location of virtually any protein within a human cell | Ztoog

    AI

    Google DeepMind’s new AI agent cracks real-world problems better than humans can

    Leave A Reply Cancel Reply

    Follow Us
    • Facebook
    • Twitter
    • Pinterest
    • Instagram
    Top Posts
    Science

    “Lensed” supernova could shed light on fundamental forces shaping Universe

    Enlarge / Gravitational lensing has revealed a beforehand unknown supernova explosion greater than 4 billion…

    AI

    Scaling audio-visual learning without labels | Ztoog

    Researchers from MIT, the MIT-IBM Watson AI Lab, IBM Research, and elsewhere have developed a…

    Crypto

    Dogecoin, XRP Beat Out Cardano, Solana To Hit New Milestone

    Kaiko, a blockchain analytics platform, performed an investigation that exposed the complexities of liquidity throughout…

    Science

    Six planets found orbiting a bright star 100 light years away

    An historical group of six exoplanets whose orbits have been unchanged for billions of years…

    Science

    The sun could contain a tiny black hole that formed in the big bang

    The sun and different stars may look and act otherwise in the event that they…

    Our Picks
    AI

    What does the future hold for generative AI? | Ztoog

    Crypto

    Bitcoin Price Is Up Despite SEC Delay, Is The Spot ETF Decision Priced In Already?

    Science

    ‘This century is special’: Astronomer Royal Martin Rees on the vast span of time

    Categories
    • AI (1,493)
    • Crypto (1,753)
    • Gadgets (1,805)
    • Mobile (1,851)
    • Science (1,866)
    • Technology (1,802)
    • The Future (1,648)
    Most Popular
    The Future

    Spain vs. England: How to Watch FIFA Women’s World Cup 2023 Final Live From Anywhere

    AI

    Region-aware pre-training for open-vocabulary object detection with vision transformers – Google Research Blog

    AI

    Generating the policy of tomorrow | Ztoog

    Ztoog
    Facebook X (Twitter) Instagram Pinterest
    • Home
    • About Us
    • Contact us
    • Privacy Policy
    • Terms & Conditions
    © 2025 Ztoog.

    Type above and press Enter to search. Press Esc to cancel.