Published May 15, 2020 | Version 1.0
Dataset Open

Audiovisual crowd counting dataset

  • 1. City University of Hong Kong
  • 2. Baidu Research
  • 3. TUM
  • 4. Northwestern Polytechnical University
  • 5. Baidu

Description

This dataset contains 1,935 annotated images, each image has one-second audio and a density map. For more details, please refer to our paper Ambient Sound Helps: Audiovisual Crowd Counting in Extreme Conditions and code.

Files

audio.zip

Files (3.7 GB)

Name Size Download all
md5:67eb05bff96e5ca396392a13c08a742d
1.4 GB Preview Download
md5:0a9633a4e190414f9c6a41aee4e06a38
122.1 MB Preview Download
md5:ab28329777d05af94eb0992d8de1da89
2.2 GB Preview Download