TokyoTech + Canon at TRECVID 2011
نویسندگان
چکیده
The aim of this section is to develop a high-performance semantic indexing system using Gaussian mixture model (GMM) supervectors and tree-structured GMMs [1, 2]. GMM spervectors corresponding to six types of audio and visual features are extracted from video shots by using tree-structured GMMs. The computational cost of maximum a posteriori (MAP) adaptation for estimating GMM parameters are reduced by tree-structured GMMs by keeping accuracy at high levels. Our best result was 17.3 % in terms of Mean InfAP, which was ranked 1st over all semantic indexing runs in the full task.
منابع مشابه
The scholarly impact of TRECVid (2003-2009)
This paper reports on an investigation into the scholarly impact of the TRECVid (TREC Video Retrieval Evaluation) benchmarking conferences between 2003 and 2009. The contribution of TRECVid to research in video retrieval is assessed by analyzing publication content to show the development of techniques and approaches over time and by analyzing publication impact through publication numbers and ...
متن کاملEvent detection: BJTU-SED at Trecvid 2011
In trecvid 2011, our team takes part in 2 event detection competition including pointing and cell to ear. We build two systems to recognize these events separately. For pointing, we extract texture and silhouette to recognize this action. And for cell to ear, we use orientation of optical flow and SVM classifier to finish this work.
متن کاملAT & T Research at TRECVID 2011
AT&T participated in two tasks at TRECVID 2011: contentbased copy detection (CCD) and instance-based search (INS). The CCD system developed for TRECVID 2010 was enhanced for speed and augmented with an additional picturein-picture detector and alternative audio features [1]. As a pilot task, participation in INS evaluated object-level contentbased copy detection and created a basis for integer-...
متن کاملBUPT - MCPRL at TRECVID 2011 *
In this paper, we describe BUPT-MCPRL systems for TRECVID 2011. Our team participated in five tasks: semantic indexing, known-item search, instance search content-based copy detection and surveillance event detection. A brief introduction is shown as follows: In this year, we proposed two different methods: one based on text and another is bio-inspired method. All 2 runs we submitted are descri...
متن کاملARTEMIS-UBIMEDIA at TRECVid 2011: Instance Search
This paper describes the approach proposed by ARTEMISUBIMEDIA team at TRECVID 2011, Instance Search (INS) task. The method is based on a semi-global image representation relying on an over-segmentation of the keyframes. An aggregation mechanism was then applied in order to group a set of sub-regions into an object similar to the query, under a global similarity criterion.
متن کامل