# Creating Realistic Character Animations with AI Networks > This patent describes how to generate lifelike digital character animations using artificial intelligence networks, specifically by predicting joint movements and overall body motion from either random inputs or specific key poses. - **Patent:** US 10964084 - **Original title:** Generating realistic animations for digital animation characters utilizing a generative adversarial network and a hip motion prediction network - **Owner:** Adobe - **Granted:** 2021 - **Status:** Active - **Times cited:** 1 - **Field:** software, ai_ml, consumer_electronics, gaming, telecommunications ## What it does The patent details a system for generating digital character animations. It starts by identifying a "code vector" (Claim 1), which can be a random input for new animations or derived from "keyframes" (Claim 2) that define specific poses. A "generative adversarial network" (GAN) then uses this code vector to create a "sequence of local poses" (Claim 1), which are the positions of a character's joints relative to its main "root joint" (like the hip). Next, a "root joint motion prediction network" calculates "hip joint velocities" and "global poses" (Claim 1), which means it figures out how the character's entire body moves in the animation space. Finally, the system generates the full animation based on these global poses. For example, an animator could provide a few key poses for a character, and the system would automatically fill in the realistic movements between them. ## What it does NOT cover - Does not cover animation generation without using a generative adversarial network (GAN) to produce local poses. - Does not cover systems that predict global body motion without specifically determining hip joint velocities from local poses. - Does not cover animation methods that rely solely on traditional manual keyframe interpolation without AI networks. - Does not cover generating animations where the root joint motion prediction network is not a convolutional neural network, as specified in Claim 3. - Does not cover systems that generate animations without first identifying a code vector or a set of keyframes as input. ## The clever bit The novelty lies in combining a generative adversarial network to create detailed local joint movements with a separate root joint motion prediction network that specifically uses hip velocities to generate realistic global body motion, ensuring the character moves naturally within the scene. ## Real-world examples 1. Adobe Character Animator 2. Video game character animation 3. Film and TV visual effects 4. Virtual reality avatars 5. Motion capture data cleanup and enhancement ## Why it matters This technology is important for making digital characters move more naturally and efficiently, reducing the manual effort for animators. It allows for the creation of diverse and realistic movements, which is crucial in video games, film, and virtual reality. Adobe, as the assignee, is a major player in creative software, and this patent reflects their investment in AI-driven content creation tools. ## Frequently asked questions ### What does Creating Realistic Character Animations with AI Networks cover? This patent describes how to generate lifelike digital character animations using artificial intelligence networks, specifically by predicting joint movements and overall body motion from either random inputs or specific key poses. ### Who owns patent US 10964084? Adobe owns this patent, granted in 2021. ### When does this patent expire? This patent is expected to expire on June 25, 2039, when the invention enters the public domain. ### What is patent US 10964084 cited by? This patent has been cited by 1 later patents that build on its ideas. ### What problem does this patent solve? This technology is important for making digital characters move more naturally and efficiently, reducing the manual effort for animators. It allows for the creation of diverse and realistic movements, which is crucial in video games, film, and virtual reality. Adobe, as the assignee, is a major player in creative software, and this patent reflects their investment in AI-driven content creation tools. ### What does this patent NOT cover? Does not cover animation generation without using a generative adversarial network (GAN) to produce local poses. **Full plain-English explainer:** https://patentbrief.org/patent/us/10964084/generating-realistic-animations-for-digital-animation-characters-utilizing-a-gen **Original patent:** https://patents.google.com/patent/US10964084 --- _Source: PatentBrief — https://patentbrief.org. Patent facts are from public records; the plain-English explanation is PatentBrief's._ ## Related patents Semantically similar inventions in the PatentBrief corpus: - [How AI Generates Images Based on Style and Content Cues](https://patentbrief.org/patent/us/12141700/generative-adversarial-network-for-generating-images) — This patent describes an artificial intelligence system that creates new images by understanding both what the image should show (like a cat) and how it should look (like a high-quality photo), using a special internal layer to blend these instructions. - [How to Train AI Models with Fake Data Using Generative Networks](https://patentbrief.org/patent/us/20230297446/data-model-generation-using-generative-adversarial-networks) — This patent describes a method for training artificial intelligence models using specially generated fake, or 'synthetic,' data created by a generative adversarial network, ensuring the synthetic data is high-quality and safe for training. - [How AI Learns to Control Game Characters Based on Their Surroundings](https://patentbrief.org/patent/us/10607134/artificially-intelligent-systems-devices-and-methods-for-learning-andor-using-an-avatars-circumstances-for-autonomous-avatar-operation) — A system that allows digital characters to automatically perform actions by matching their current environment to previously learned experiences stored in a database. - [How AI Transfers Visual Styles Between Images While Keeping Structure](https://patentbrief.org/patent/us/12282696/method-and-system-for-semantic-appearance-transfer-using-splicing-vit-features) — This patent describes an AI method to generate a new image that combines the structure of one input image with the visual style of another, semantically related image, using a pre-trained Vision Transformer model. - [How Cloud Systems Automatically Create and Train AI Data Models](https://patentbrief.org/patent/us/11615208/dall-e-text-to-image-generation) — A cloud-based system that generates fake, privacy-safe data to train AI models, ensuring they remain accurate while protecting sensitive personal information.