TUKE MediaEval 2012: Spoken Web Search using DTW and Unsupervised SVM
نویسندگان
چکیده
This working paper provides the basic information about experiments conducted on audio documents within the MediaEval 2012 spoken web search evaluation project. The main purpose of these experiments was to build a robust and language independent system for spoken term detection. Therefore we have proposed query-by-example searching system based on the minimum-cost alignment of DTW algorithm and unsupervised SVM misclassification rate. Results show that our system is liable to variable length of queries with similar spectral characteristics that results in poor detection performance with high number of insertions and misdetections. There were no other resources used during the development.
منابع مشابه
TUKE at MediaEval 2015 QUESST
In this paper, we present our retrieving system for QUery by Example Search on Speech Task (QUESST), comprising the posteriorgram-based modeling approach along with the weighted fast sequential dynamic time warping algorithm (WFS-DTW). For this year, our main effort was directed toward developing language-dependent keyword matching system, utilizing all available information about spoken langua...
متن کاملThe JHU-HLTCOE Spoken Web Search System for MediaEval 2012
We present an overview for a truly zero resource query-byexample search system designed for the 2012 MediaEval Spoken Web Search task. Our system is based on the recently proposed randomized acoustic indexing and logarithmictime search (RAILS) framework. The input is merely the raw acoustic observations for the query and search collection, requiring no trained models whatsoever, not even unsupe...
متن کاملThe TUM Cumulative DTW Approach for the Mediaeval 2012 Spoken Web Search Task
This paper describes the system proposed for the Spoken Web Search task at Mediaeval 2012 campaign. We use an audio-only system based on our new called Cumulative Dynamic Time Warping (CDTW) algorithm. This algorithm combines the scores of all the alignment paths and allows for the learning of different cost functions for each kind of step in the alignment matrix (diagonal, horizontal and verti...
متن کاملBUT2012 Approaches for Spoken Web Search - MediaEval 2012
We submitted two approaches as the required runs: Acoustic Keyword Spotting as the primary one (AKWS) and Dynamic Time Wrapping as the secondary one (DTW) for the Spoken Web Search task. We aimed at building a simple phone based language-dependent system. We experimented with universal context bottle-neck neural network classifier with 3-state phone posterior features or bottle-neck features.
متن کاملTUKE System for MediaEval 2014 QUESST
Two approaches to QbE (Query-by-Example) retrieving system, proposed by the Technical University of Košice (TUKE) for the query by example search on speech task (QUESST), are presented in this paper. Our main interest was focused on building such QbE system, which is able to retrieve all given queries with and without using any external speech resources. Therefore we developed posteriorgram-bas...
متن کامل