Dear all, We are pleased to release concept detection scores for TRECVID-2009 test data using the CU-VIREO374 models. These models were trained on the TRECVID-2005 data set (338 concepts) and the TRECVID-2007 data set (36 concepts), using both local and global features. In addition, this release also includes detection scores of the 20 concepts announced this year (for the HLF task), generated from *new* models trained on *TRECVID-2009 development data* using the collaborative annotations. For the overlapped concepts (~10) between the 374 set and this year's 20 concepts, detection scores from both the old and new models are available. Please visit the link below for download instructions and relevant reference: http://www.ee.columbia.edu/ln/dvmm/CU-VIREO374/ Yu-Gang Jiang, Akira Yanagawa, Shih-Fu Chang, Chong-Wah Ngo, "CU-VIREO374: Fusing Columbia374 and VIREO374 for Large Scale Semantic Concept Detection", Columbia University ADVENT Technical Report #223-2008-1, Aug. 2008 Please feel free to send me a message if you have any further questions. Best regards, Yu-Gang Jiang