ADVANCE (AuDio Visual Aerial sceNe reCognition datasEt)

Audio Visual Aerial 场景识别数据集（ADVANCE）是一个全新的多模态学习数据集，旨在探索音频和传统视觉信息对场景识别的贡献。该数据集共包含 5075 对带地理标记的航拍图像和声音，分为 13 个场景类别，即机场、运动场、海滩、桥梁、农田、森林、草原、港口、湖泊、果园、住宅区、灌木地和火车站。

Audio-Visual Aerial Scene Recognition Dataset (ADVANCE) is a novel multimodal learning dataset designed to explore the contributions of audio and conventional visual information to scene recognition. This dataset contains a total of 5075 pairs of geotagged aerial images and audio recordings, which are divided into 13 scene categories, namely airports, sports fields, beaches, bridges, farmlands, forests, grasslands, ports, lakes, orchards, residential areas, shrublands, and railway stations.

搜集汇总

背景与挑战

背景概述

ADVANCE是一个多模态航拍场景识别数据集，包含5075对带地理标记的航拍图像和声音，覆盖13个不同场景类别。该数据集由慕尼黑工业大学、百度等机构联合发布，旨在探索音频和视觉信息在场景识别中的联合作用。