Kinect database. Foreground segmentation

Resource Type Date
Dataset 2013-01-02

Description

Nine single person sequences, recorded with a kinect device, to show depth and color camouflage situations that are prone to errors in color-depth scenarios.

Nine video sequences can be previewed/downloaded from the following links:

 File name and link Properties Description

office1.avi, 4.9M

Visual and depth side-by-side

1280x480@3fps, length 00:50
Codec MS MPEG-4 Video v2 (MP42)
Decoded: Planar 4:2:0 YUV

Person1 in office. Starts with 13sec scene background
office2.avi, 5.8M

Visual and depth side-by-side

1280x480@3fps, length 01:00
Codec MS MPEG-4 Video v2 (MP42)
Decoded: Planar 4:2:0 YUV

Person2 in office. Starts with 16sec scene background
sroom1.avi, 12M

Visual

640x480@10fps, length 00:50
Codec MS MPEG-4 Video v2 (MP42)
Decoded: Planar 4:2:0 YUV

Person1 in smart room. No chairs...
sroom2.avi, 14M  ...  One chair...
sroom3.avi, 12M    No chairs...
sroom4.avi, 18M    Two chairs...
sroom5.avi, 16M    
sroom6.avi, 17M    
sroom7.avi, 8.6M    

People involved

Jaume Gallego PhD Candidate
Montse Pardàs Professor

Related Publications