• What Was Once Sci-Fi is Now Reality!
    WARNING: GRAPHIC CONTENT:
    Don't play until the end to avoid graphic content from The Running Man film.

    This is SPARK, a brand-new AI model that can create 3D face reconstructions from multiple, unconstrained portrait videos and enable real-time tracking on unseen footage. It’s the result of a collaboration between Technicolor, Inria, and the Max Planck Institute.

    SPARK allows the creation of a relightable face avatar from multiple monocular videos—bringing sci-fi-like technology into the real world.

    What makes SPARK truly special is its precise face tracking, which can be used for face editing and visual effects IN REAL-TIME!
    Tasks that would typically require manual 3D modeling or complex scanning sessions of actors can now be automated!

    This technology opens up endless possibilities in professional fields like FX, virtual production, 3D modeling, and gaming, but is not exempt of many ethical concerns.

    This time, the potential for misuse, especially in altering real-time footage for fake news and deep fakes, is completely real in the wrong hands.

    As a fan of sci-fi and dystopian films, many of them have warned us about its risks. I specially remember "The Running Man", where they faked the death of Ben Richards by overlaying his face in real time with what the called "Digital Matte Tracking"

    Link to the model paper and samples in comments!
    Code is not yet available!

    Content Credits:
    © The Running Man, 1987.
    TAFT Entertainment Pictures / Keith Barish Productions.

    #ArtificialIntelligence #Innovation#Future
    🤯 What Was Once Sci-Fi is Now Reality! ⚠️ WARNING: GRAPHIC CONTENT: Don't play until the end to avoid graphic content from The Running Man film. This is SPARK, a brand-new AI model that can create 3D face reconstructions from multiple, unconstrained portrait videos and enable real-time tracking on unseen footage. It’s the result of a collaboration between Technicolor, Inria, and the Max Planck Institute. SPARK allows the creation of a relightable face avatar from multiple monocular videos—bringing sci-fi-like technology into the real world. What makes SPARK truly special is its precise face tracking, which can be used for face editing and visual effects IN REAL-TIME! Tasks that would typically require manual 3D modeling or complex scanning sessions of actors can now be automated! 🔝 This technology opens up endless possibilities in professional fields like FX, virtual production, 3D modeling, and gaming, but is not exempt of many ethical concerns. 🚩 This time, the potential for misuse, especially in altering real-time footage for fake news and deep fakes, is completely real in the wrong hands. ⚠️ As a fan of sci-fi and dystopian films, many of them have warned us about its risks. I specially remember "The Running Man", where they faked the death of Ben Richards by overlaying his face in real time with what the called "Digital Matte Tracking" 📌 Link to the model paper and samples in comments! Code is not yet available! Content Credits: © The Running Man, 1987. TAFT Entertainment Pictures / Keith Barish Productions. #ArtificialIntelligence #Innovation#Future
    Wow
    1
    0 Comentários 1 Compartilhamentos 3K Visualizações 4
  • Unity's AI-based platform for game development, Unity Muse, has been updated and can now generate better textures.

    #unity #madewithunity #unityengine #unitymuse #gamedev #indiedev #gamedevelopment #indiegamedev #ai #artificialintelligence #generativeai #genai #aitech #aitechnology

    Unity's AI-based platform for game development, Unity Muse, has been updated and can now generate better textures. #unity #madewithunity #unityengine #unitymuse #gamedev #indiedev #gamedevelopment #indiegamedev #ai #artificialintelligence #generativeai #genai #aitech #aitechnology
    Yay
    1
    0 Comentários 0 Compartilhamentos 9K Visualizações
  • A concept of a cooking app in AR

    How could AR devices help people cook? Software engineer Stijn Spanhove shared a cool experiment with a multimodal LLM in WebAR. The demo shows AI identifying ingredients, setting up bounding boxes, and preparing them for AR. The app then also generates recipes and corresponding images.
    #xr #ar #AI #artificialintelligence #augementedreality
    A concept of a cooking app in AR 👨‍🍳 How could AR devices help people cook? Software engineer Stijn Spanhove shared a cool experiment with a multimodal LLM in WebAR. The demo shows AI identifying ingredients, setting up bounding boxes, and preparing them for AR. The app then also generates recipes and corresponding images. #xr #ar #AI #artificialintelligence #augementedreality
    Like
    1
    0 Comentários 0 Compartilhamentos 4K Visualizações 16
  • Controlling multiple areas of your video generations with AI

    Runway presented the Multi Motion Brush in Gen-2, its multimodal AI system that generates videos using text, images, and video clips. Basically, it's an update to a brush that can animate elements in static images, but this time you can manipulate multiple objects simultaneously.



    #runway #AI #artificialintelligence #ML
    Controlling multiple areas of your video generations with AI 🦅 Runway presented the Multi Motion Brush in Gen-2, its multimodal AI system that generates videos using text, images, and video clips. Basically, it's an update to a brush that can animate elements in static images, but this time you can manipulate multiple objects simultaneously. #runway #AI #artificialintelligence #ML
    0 Comentários 0 Compartilhamentos 4K Visualizações 15
  • Controlling multiple areas of your video generations with AI

    #Runway presented the Multi Motion Brush in Gen-2, its multimodal AI system that generates videos using text, images, and video clips. Basically, it's an update to a brush that can animate elements in static images, but this time you can manipulate multiple objects simultaneously.



    #runway #AI #artificialintelligence #ML
    Controlling multiple areas of your video generations with AI 🦅 #Runway presented the Multi Motion Brush in Gen-2, its multimodal AI system that generates videos using text, images, and video clips. Basically, it's an update to a brush that can animate elements in static images, but this time you can manipulate multiple objects simultaneously. #runway #AI #artificialintelligence #ML
    0 Comentários 0 Compartilhamentos 4K Visualizações 18