Recent Extensions of VIREO-374

-- Features and prediction scores on recent datasets

Yu-Gang Jiang¹, Chong-Wah Ngo¹ and Jun Yang²

¹Video Retrieval Group (VIREO), City University of Hong Kong, Hong Kong
²School of Computer Science, Carnegie Mellon University, USA
{yjiang, cwngo}@cs.cityu.edu.hk; juny@cs.cmu.edu

This page contains features (color, texture, and keypoint-based) and the VIREO-374 concept detection scores (SVM predictions) on several recent datasets.

Citation of VIREO-374:

Yu-Gang Jiang, Jun Yang, Chong-Wah Ngo, Alexander G. Hauptmann
Representations of Keypoint-Based Semantic Concept Detection: A Comprehensive Study
IEEE Transactions on Multimedia, 2010.
Yu-Gang Jiang, Chong-Wah Ngo, Jun Yang
Towards Optimal Bag-of-Features for Object Categorization and Semantic Video Retrieval
ACM International Conference on Image and Video Retrieval (CIVR'07), Amsterdam, The Netherlands, 2007.

NEW! TRECVID 2009 Dataset:

Download Detection Scores and Features:

Please note that a copy of the same detection scores is also placed in the CU-VIREO374 download site. In CU-VIREO374 we also provide additional detection scores for the 20 new concepts announced in TRECVID-2009, generated from our new models trained on TRECVID-2009 development set.

To download the detection scores and features computed on the TRECVID-2009 test data, please fill out the following form. We will send you the download instructions via email immediately.

TRECVID 2008 Dataset:

Download Detection Scores: Jointly released with Columbia374, please visit CU-VIREO374 download site.

Download Features:

To download the features of TRECVID-2008 Data, please fill out the following form. We will send you the download instructions via email immediately.

TRECVID 2007 Dataset:

Download Detection Scores and Features:

Please fill out the following form. We will send you the download instructions via email immediately.

Top 200 ranked keyframes (shots) in TRECVID-07 test set:

Last updated by Yu-Gang Jiang on Aug., 2008.