Close Menu
Ztoog
    What's Hot
    Science

    Quantum time travel: The experiment to ‘send a particle into the past’

    Technology

    SmileDirectClub Shut Down: What We Know About Payments and Finding New Treatment

    The Future

    AIs trained on AI-generated images produce glitches and blurs

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

      OPPO launches A5 Pro 5G: Premium features at a budget price

      How I Turn Unstructured PDFs into Revenue-Ready Spreadsheets

      Is it the best tool for 2025?

      The clocks that helped define time from London’s Royal Observatory

      Summer Movies Are Here, and So Are the New Popcorn Buckets

    • Technology

      What It Is and Why It Matters—Part 1 – O’Reilly

      Ensure Hard Work Is Recognized With These 3 Steps

      Cicada map 2025: Where will Brood XIV cicadas emerge this spring?

      Is Duolingo the face of an AI jobs crisis?

      The US DOD transfers its AI-based Open Price Exploration for National Security program to nonprofit Critical Minerals Forum to boost Western supply deals (Ernest Scheyder/Reuters)

    • Gadgets

      Maono Caster G1 Neo & PD200X Review: Budget Streaming Gear for Aspiring Creators

      Apple plans to split iPhone 18 launch into two phases in 2026

      Upgrade your desk to Starfleet status with this $95 USB-C hub

      37 Best Graduation Gift Ideas (2025): For College Grads

      Backblaze responds to claims of “sham accounting,” customer backups at risk

    • Mobile

      Motorola’s Moto Watch needs to start living up to the brand name

      Samsung Galaxy S25 Edge promo materials leak

      What are people doing with those free T-Mobile lines? Way more than you’d expect

      Samsung doesn’t want budget Galaxy phones to use exclusive AI features

      COROS’s charging adapter is a neat solution to the smartwatch charging cable problem

    • Science

      Nothing is stronger than quantum connections – and now we know why

      Failed Soviet probe will soon crash to Earth – and we don’t know where

      Trump administration cuts off all future federal funding to Harvard

      Does kissing spread gluten? New research offers a clue.

      Why Balcony Solar Panels Haven’t Taken Off in the US

    • AI

      Hybrid AI model crafts smooth, high-quality videos in seconds | Ztoog

      How to build a better AI benchmark

      Q&A: A roadmap for revolutionizing health care through data-driven innovation | Ztoog

      This data set helps researchers spot harmful stereotypes in LLMs

      Making AI models more trustworthy for high-stakes settings | Ztoog

    • Crypto

      Ethereum Breaks Key Resistance In One Massive Move – Higher High Confirms Momentum

      ‘The Big Short’ Coming For Bitcoin? Why BTC Will Clear $110,000

      Bitcoin Holds Above $95K Despite Weak Blockchain Activity — Analytics Firm Explains Why

      eToro eyes US IPO launch as early as next week amid easing concerns over Trump’s tariffs

      Cardano ‘Looks Dope,’ Analyst Predicts Big Move Soon

    Ztoog
    Home » University of Cambridge Researchers Introduce a Dataset of 50,000 Synthetic and Photorealistic Foot Images along with a Novel AI Library for Foot
    AI

    University of Cambridge Researchers Introduce a Dataset of 50,000 Synthetic and Photorealistic Foot Images along with a Novel AI Library for Foot

    Facebook Twitter Pinterest WhatsApp
    University of Cambridge Researchers Introduce a Dataset of 50,000 Synthetic and Photorealistic Foot Images along with a Novel AI Library for Foot
    Share
    Facebook Twitter LinkedIn Pinterest WhatsApp

    The well being, vogue, and health industries are extremely within the tough laptop imaginative and prescient downside of 3D reconstructing human physique components from footage. They deal with the problem of reconstructing a human foot on this examine. Accurate foot fashions are helpful for shoe procuring, orthotics, and private well being monitoring, and the concept of recovering a 3D foot mannequin from footage has develop into extremely engaging because the digital market for these companies grows. There are 4 varieties of current foot reconstruction options: Costly scanning equipment is one technique reconstruction of noisy level clouds, utilizing depth maps or phone-based sensors like a TrueDepth digital camera, is one other Structure from Motion (SfM) it’s adopted by Multi-View Stereo (MVS) and generative foot fashions are fitted to image silhouettes is a fourth technique. 

    They conclude that none of these choices is ample for exact scanning in a home setting: Most individuals can not afford costly scanning tools; phone-based sensors will not be extensively accessible or user-friendly; noisy level clouds are difficult to make the most of for actions that come after, such rendering and measuring; Additionally, foot generative fashions have been low high quality and restrictive, and utilizing solely silhouettes from photographs limits the quantity of geometrical info that may be obtained from the photographs, which is particularly problematic in a few-view setting. SfM will depend on many enter views to match dense options between photographs, and MVS can even produce noisy level clouds. 

    The inadequate availability of paired footage and 3D floor fact knowledge for toes for coaching additional constrains the efficiency of these approaches. To do that, researchers from the University of Cambridge current FOUND, or Foot Optimisation, utilizing Uncertain Normals for Surface Deformation. This algorithm makes use of uncertainties along with per-pixel floor normals to enhance upon standard multi-view reconstruction optimization approaches. Like, their approach wants a minimal quantity of enter RGB pictures which were calibrated. Despite relying simply on silhouettes, that are devoid of geometric info, they use floor normals and key factors as supplementary clues. They additionally make accessible a sizable assortment of artificially photorealistic photographs matched with floor fact labels for these sorts of alerts to beat knowledge shortage. 

    Their essential contributions are outlined under: 

    • They launch SynFoot, a large-scale artificial dataset of 50,000 photorealistic foot footage with exact silhouettes, floor regular, and keypoint labels, to help in analysis on 3D foot reconstruction. Although acquiring such info on precise photographs necessitates expensive scanning equipment, their dataset displays nice scalability. They exhibit that their artificial dataset captures sufficient variance inside foot footage for downstream duties to generalize to actual photographs regardless of solely having 8 real-world foot scans. Additionally, they make accessible an analysis dataset consisting of 474 photographs of 14 precise toes. Each matched with high-resolution 3D scans and ground-truth per-pixel floor normals. Lastly, they make identified their proprietary Python library for Blender, which permits for the efficient creation of large-scale artificial datasets. 

    • They present that an uncertainty-aware floor regular estimate community can generalize to precise in-wild foot footage after coaching solely on their artificial knowledge from 8 foot scans. To cut back the distinction within the area between synthetic and genuine foot photographs, they make use of aggressive look and perspective augmentation. The community calculates the related uncertainty and floor normals at every pixel. The uncertainty is useful in two methods: first, by thresholding the uncertainty, they’ll receive exact silhouettes with out having to coach a totally different community; second, through the use of the estimated uncertainty to weight the floor regular loss of their optimization scheme, they’ll improve robustness towards the chance that the predictions made in some views might not be correct. 

    • They present an optimization technique that makes use of differentiable rendering to suit a generative foot mannequin to a sequence of calibrated photographs with anticipated floor normals and key factors. Their pipeline outperforms state-of-the-art photogrammetry for floor reconstruction, is uncertainty-aware, and can rebuild a watertight mesh from a restricted quantity of views. It may also be used for knowledge obtained from a shopper’s mobile phone.


    Check out the Paper and Project. All credit score for this analysis goes to the researchers of this venture. Also, don’t overlook to hitch our 32k+ ML SubReddit, 40k+ Facebook Community, Discord Channel, and Email Newsletter, the place we share the newest AI analysis information, cool AI tasks, and extra.

    If you want our work, you’ll love our e-newsletter..

    We are additionally on Telegram and WhatsApp.


    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 aimed toward harnessing the ability of machine studying. His analysis curiosity is picture processing and is enthusiastic about constructing options round it. He loves to attach with individuals and collaborate on fascinating tasks.


    🔥 Meet Retouch4me: A Family of Artificial Intelligence-Powered Plug-Ins for Photography Retouching

    Share. Facebook Twitter Pinterest LinkedIn WhatsApp

    Related Posts

    AI

    Hybrid AI model crafts smooth, high-quality videos in seconds | Ztoog

    AI

    How to build a better AI benchmark

    AI

    Q&A: A roadmap for revolutionizing health care through data-driven innovation | Ztoog

    AI

    This data set helps researchers spot harmful stereotypes in LLMs

    AI

    Making AI models more trustworthy for high-stakes settings | Ztoog

    AI

    The AI Hype Index: AI agent cyberattacks, racing robots, and musical models

    AI

    Novel method detects microbial contamination in cell cultures | Ztoog

    AI

    Seeing AI as a collaborator, not a creator

    Leave A Reply Cancel Reply

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

    Black Friday: Get a great deal on Galaxy Watch6, iPad 10.9″ 2022 and more in Germany

    Here are all the perfect Black Friday offers on smartwatches and tablets from Amazon Germany.…

    Technology

    A new generation of storm chasers takes on Mother Nature in Twisters trailer

    Daisy Edgar-Jones, Anthony Ramos, and Glen Powell star in Twisters, a standalone movie impressed by…

    AI

    Is Multilingual AI Truly Safe? Exposing the Vulnerabilities of Large Language Models in Low-Resource Languages

    GPT-4 defaults to saying, “Sorry, but I can’t help with that,” in reply to requests…

    Crypto

    Sam Bankman-Fried says he didn’t defraud FTX customers

    He stated that Alameda “borrowed” them On Friday, Sam Bankman-Fried sat on the stand, as…

    The Future

    Apple CEO reaches out to Beijing amid iPhone woes

    Apple CEO Tim Cook this week launched into a go to to China to strengthen…

    Our Picks
    Mobile

    What is a 0.5 selfie and how can you create one?

    The Future

    The best apps are taking a page from mobile gaming

    Crypto

    Ripple Token Unfazed By Crypto Turmoil With 60% Rally

    Categories
    • AI (1,483)
    • Crypto (1,745)
    • Gadgets (1,796)
    • Mobile (1,840)
    • Science (1,854)
    • Technology (1,790)
    • The Future (1,636)
    Most Popular
    Mobile

    How to create your own chatbot with OpenAI custom GPTs

    Science

    Fukushima water releases prompt debate over tritium in seafood

    Crypto

    Bitcoin Investors Place $500 Million In Spot BTC Bids Below Market Price

    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.