Structural Scene Image Dataset (SSDB40)

SSDB40 dataset was applied to the image segmentation experiment for structural scenes in the article:
[Lin et al. 2015] I.-C. Lin, Y.-C. Lan, P.-W. Cheng, "SI-Cut: Structural Inconsistency Analysis for Image Foreground Extraction," IEEE Trans. Visualization and Computer Graphics, 21(7):860-872, July, 2015. 

\Image
The sources of these 40 images were selected from the LabelMe webpage: http://labelme.csail.mit.edu .
The credits for the image sources go to the LabeMe database: 
[Russel et al. 2008] B. Russell, A. Torralba, K. Murphy, W. T. Freeman, "LabelMe: a database and web-based tool for image annotation," Intl. J. Computer Vision, 77(1-3): pp 157-173, 2008.

\BoundingBox
The binary maskes show the indicated rectangles used in the structural-scene experiment in [Lin et al. 2015].

\GroundTruth
A large portion of ground-truth masks in LabelMe [Russel et al. 2008] are approximated by polygonal contours. 
For more accurate experiments, the ground-truth masks were refined by users and used in  [Lin et al. 2015].









