NYU AI Spring 2026 — Assignment 2
This repository contains an object-detection database (index) generated from a video using a YOLO-based detection pipeline.The dataset enables image-to-video semantic retrieval by matching detected object components between query images and indexed video frames.
Video processed: Toyota RAV4 review video
Frame sampling rate: 1 FPS
Detection model: YOLOv8 (fine-tuned car-parts model)
Detector… See the full description on the dataset page:
https://huggingface.co/datasets/simonhu125/rav4-detections.
