README.md

---
title: Drone Tracking Assignment AI Spring26
emoji: 🐢
colorFrom: yellow
colorTo: red
sdk: gradio
sdk_version: 6.9.0
app_file: app.py
pinned: false
---

# UAV Drone Detection Frames

This dataset contains frames with at least one drone detection, stored in Parquet format.

## Features

- `video`: source video name
- `frame_file`: frame filename
- `image`: embedded image
- `objects.bbox`: bounding boxes in `[x, y, width, height]`
- `objects.categories`: class ids
- `scores`: detection confidence scores

## Description

The dataset was created by extracting frames from drone videos, running a fine-tuned YOLO detector, and keeping only frames with at least one detection.

This dataset was prepared for an academic drone detection and tracking assignment at NYU for the Intro to AI class Spring 2026.