The app combines the power of Gemini ⚡ with the speed of a local tracker. It first uses Gemini’s capabilities to locate a specific object 🎯. While the initial request may take a moment ⏳, once the object is found, the app switches to a local tracker (OpenCV MIL) for real-time tracking, ensuring instant and efficient feedback. Users are kept informed of progress through a simple beeping sound 🔊, making the experience both intuitive and seamless.

I can ask the app for something more complex for a basic image detector, for example to find a ripe tomato. 🍅

Presentation and demo

I entered with this project in the Google Gemini competition, you can vote it here🙂