This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass Autumn Hunt AI is a scavenger hunt that encourages users to get outside and snap some photos of autumn. This web app uses local AI (via Transformers.js) to detect autumn objects. The user snaps photos, which are analyzed locally on the user's phone, for autumn objects on their scavenger hunt list. As objects are detected, they are automatically crossed off the list.
https://github.com/RyanCodeDev/autumn-hunt-ai This web app was built in React using Vite. A small local object detection model (onnx-community/rtdetr_v2_r50vd-ONNX) is loaded in the browser on the user's device. As the user snaps photos, the image is analyzed by the local model to detect objects in the image. If the object matches one of the items on the user's scavenger hunt list (or matches a predefined variation of its name), it is automatically crossed off the user's list.
This web app is completely private as no images leaves the user device. All image processing happens on device.