Description

The crowd datasets are collected from a variety of sources, such as UCF and data-driven crowd datasets. The sequences are diverse, representing dense crowd in the public spaces in various scenarios such as pilgrimage, station, marathon, rallies and stadium. In addition, the sequences have different field of views, resolutions, and exhibit a multitude of motion behaviors that cover both the obvious and subtle instabilities. We annotate the datasets manually in order to ease researcher to evaluate their respective framework. It has been used in the paper: M.K. Lim, V.J. Kok, C.C. Loy and C.S. Chan "Crowd Saliency Detection via Global Similarity Structure ", in ICPR 2014. There are two(2) folders associated with the dataset and a ReadMe file: - Crowd Sequence Folder name = Crowd Sequence Total Sequences = 20 -Dataset Groundtruth Folder name = Dataset Groundtruth Total Files (text) = 20 The salient regions are defined by bounding box information in the form of either: * Rectangle Properties (Frame no; Region Identifier; Minimum x-coordinate; Minimum y-coordinate; Maximum x-coordinate; Maximum y-coordinate) * Example: 34 1 75 76 140 130 * Polygon Properties (Frame no; Region Identifier; Point 1 x-coordinate; Point 1 y-coordinate; Point 2 x-coordinate; Point 2 y-coordinate; Point 3 x-coordinate; Point 3 y-coordinate; Point 4 x-coordinate; Point 4 y-coordinate) * Example: 1 11 153 182 323 121 342 173 170 232 If you use this dataset in your work, you should reference: M.K. Lim, V.J. Kok, C.C. Loy and C.S. Chan "Crowd Saliency Detection via Global Similarity Structure ", in ICPR 2014, pp. 3957-3962. @inproceedings{DBLP:conf/icpr/LimKLC14, author = {Mei Kuan Lim and Ven Jyn Kok and Chen Change Loy and Chee Seng Chan}, title = {Crowd Saliency Detection via Global Similarity Structure}, booktitle = {22nd International Conference on Pattern Recognition, {ICPR} 2014, Stockholm, Sweden, August 24-28, 2014}, pages = {3957--3962}, year = {2014}, crossref = {DBLP:conf/icpr/2014}, url = {http://dx.doi.org/10.1109/ICPR.2014.678}, doi = {10.1109/ICPR.2014.678}, }

Related datasets