TRECVID 2011 - An Overview of the Goals, Tasks, Data,Evaluation Mechanisms, and Metrics

By Paul Over, George Awad, Jonathan G. Fiscus, Brian Antonishek, Martial Michel, Wessel Kraaij, Alan F. Smeaton, Georges Quénot. 2011

485 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

1. Yi Yang
2. Dacheng Tao
3. Zhe Gan
4. Luc Van Gool
5. Trevor Darrell
6. Mohit Bansal
7. Limin Wang
8. Abhinav Kumar Gupta

Citations in the corpus

• Every Moment Counts: Dense Detailed Labeling of Actions in Complex Videos (2015, neutral): “Zha et al. [40] examine aggregation strategies for combining deep learned image-based features for each frame, obtaining impressive results on TRECVID MED retrieval.”
• Hierarchical Deep Temporal Models for Group Activity Recognition (2016, background): “In the context of surveillance video, the TRECVid Surveillance Event Detection [54], UT-Interaction [55], VIRAT [56], and UCLA Courtyard datasets [57] are examples of challenging tasks including individual and pair-wise interactions.”
• Blind late fusion in multimedia event retrieval (2016, methodology): “In our experiments, we use simulations and a real-world multimedia event retrieval data set named TRECVID MED [20] to find out in which fusion method performs best under which circumstances.”
• Large-scale visual sentiment ontology and detectors using adjective noun pairs (2013, methodology): “It would be useful to check the performance degradation if we apply SentiBank detectors to new data domains, such as TRECVID or PASCAL VOC.”
• Soccer video and player position dataset (2014, background): “A multitude of datasets for various purposes in multimedia research do already exists, like the TrecVid collection where both videos and ground truths are provided [9].”
• Compositional Models for Video Event Detection: A Multiple Kernel Learning Latent Variable Approach (2013, methodology): “We evaluate our model on the challenging TRECVID MED11 dataset [9], following a standard evaluation protocol used in previous work [12].”
• The YLI-MED Corpus: Characteristics, Procedures, and Plans (2015, methodology): “Table 1: The fifteen events for the TRECVID MED 2011 Evaluation (Over et al. 2012).”
• A Survey of Zero-Shot Learning (2019, methodology): “For video retrieval (event detection) problems, the TRECVID dataset [112] is widely used.”
• The ImageNet Shuffle: Reorganized Pre-training for Video Event Detection (2016, methodology): “The TRECVID Multimedia Event Detection 2013 dataset consists of roughly 27,000 test videos [17].”
• Video Event Detection by Inferring Temporal Instance Labels (2014, methodology): “To evaluate our framework, we conducted experiments on three large-scale video datasets: TRECVID Multimedia Event Detection (MED) 2011, MED 2012 [15] and Columbia Consumer Videos (CCV) [10] datasets.”