# BinEgo‑360 Challenge - Classification - Extend — Data

来源：https://www.kaggle.com/competitions/bin-ego-360-challenge-classification-ext/data

本地 Meta Kaggle 快照；不是本次实时页面。

# Dataset Description

The challenge uses our public dataset 360+x for training/validation, and a held-out test set for the evaluation. The full dataset can be fetched: [https://huggingface.co/datasets/quchenyuan/360x_dataset_HR](https://huggingface.co/datasets/quchenyuan/360x_dataset_HR)


![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F16342273%2F645b624e0405ed841981d6377c0e9498%2Fteaser_video.gif?generation=1748609310125434&alt=media)

Dataset Overview:
360+x captures real-world environments from four synchronised viewpoints (360° panoramic, binocular egocentric, monocular egocentric, third-person front) together with 6-channel spatial audio, GPS, weather and text metadata. The dataset is designed to study how an embodied agent could perceive, localise actions and recognise scenes when immersed in a space, while also observing itself from an external vantage.