gemini Gemini Blog ·

Google showcases 5 developer projects built with Gemini Omni Flash

aiengineer
announcement

Google details a selection of five innovative projects developers have created using Gemini Omni Flash. This post showcases various applications of the AI model, which was introduced at I/O and recently made available to developers. Builders are leveraging its capabilities for video editing, scene transformation, object animation, and style application. The examples demonstrate Omni's power in generating and editing realistic video content from diverse inputs.

  • Omni for Multi-Perspective Video Editing
  • Omni for Dynamic Scene Transformation
  • Omni in Google Flow for Animating Sketches
  • Omni for Video Style and Effect Application
  • Omni for Visualizing Abstract Concepts
Notes (5)
  • Omni for Multi-Perspective Video Editing

    Builder Leon Lin demonstrated Gemini Omni Flash's ability to change camera angles, environments, and apply cinematic zooms within a video without losing scene coherence. The example showed capturing a woman from about 20 different perspectives with shifting backgrounds.

  • Omni for Dynamic Scene Transformation

    Carlos Santana showcased how Gemini Omni Flash can alter elements within a video, such as changing lighting from day to night, adding weather effects, and modifying foliage color. The model maintains a coherent scene despite these significant changes.

  • Omni in Google Flow for Animating Sketches

    Pan demonstrated using Gemini Omni Flash in Google Flow to transform sketches into realistic videos. Doodles guided the movement of individual elements, turning everyday objects like a lemon or espresso cup into animated scenes.

  • Omni for Video Style and Effect Application

    Jerrod Lew used Gemini Omni Flash in Google Flow to render a video in various animation styles, including live-action, anime, and claymation. The model blends user-defined aesthetics and natural language descriptions to create cohesive clips.

  • Omni for Visualizing Abstract Concepts

    The Hyperagent team utilized Gemini Omni Flash to visualize concepts, layering landscaping into an empty park video, personifying data with an animated professor, and gamifying a to-do list. Omni helps create scenes following real-world logic.

Read the original announcement →

https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-omni-builders/

Related releases