Jumas @ Trecvid 2010
نویسندگان
چکیده
We summarize the fully automatic approach to the TRECVID 2010 Known Item (KIS) and Instance Search (INS) tasks of the budapest_acad team in the JUMAS Consortium with Fondazione Bruno Kessler who provided the ASR technologies. Our submissions summarized in Table 1 use linear combinations of the following basic techniques. • Meta: Text retrieval based on video metadata. • ASR: Text retrieval based on ASR most likely readout. • SIX and CLEF: Total weight of high level feature classifiers considered relevant by text based similarity to the topic. We used the publicly available Semantic Indexing (SIX) category predictions and the ImageCLEF annotations (CLEF). Our KIS submissions were based on the linear combination of the four components. Based on experiments from last year, we expected SIX to perform best and gave high weight in our runs. This assumption proved to be wrong. It turns out that Meta contained the strongest information with little improvement possible beyond based on simple linear combination. For Instance Search we submitted a single run where we used the Dutch translation of the queries to retrieve the ASR text. This run reached an AP sum of 0.508 for the 22 topics. The INS task proved to be very difficult. Our fairly straightforward method reached close to the 0.729 result of the best participant.
منابع مشابه
The Scholarly Impact of TRECVid ( 2003 - 2009 ) ( pre - print )
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 کاملJOANNEUM RESEARCH and Vienna University of Technology at TRECVID 2010
We participated in two tasks: semantic indexing (SIN) and instance search (INS).
متن کاملYork_University at TRECVID 2010
In this paper, we describe our work done by members at York University in Canada for the KIS (Known-item search) task of TRECVID 2010. This is the first time that we participate in the TRECVID. With rich experience in text retrieval, we mainly focus on the meta information of videos, and try to figure out the importance of these description corpus. In order to obtain this goal, we do not use an...
متن کاملAT&T Research at TRECVID 2010
AT&T participated in two tasks at TRECVID 2010: contentbased copy detection (CCD) and instance-based search (INS). The CCD system developed for TRECVID 2009 was enhanced for efficiency and scale and was augmented by audio features [1]. As a pilot task, participation in INS was meant to evaluate a number of algorithms traditionally used for search in a fully automated setting. In this paper, we ...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2010