The Spoken Web Search task at MediaEval 2011

Florian Metze (1) , Nitendra Rajput (2) , Xavier Anguera (3) , Marelie Davel (4) , Guillaume Gravier (5) , Charl van Heerden (4) , Gautam Mantena (6) , Armando Muscariello (7) , Kishore Pradhallad (6) , Igor Szöke (8) , Javier Tejedor (9)

1 CMU - Carnegie Mellon University [Pittsburgh]
2 IBM Almaden Research Center [San Jose]
3 Telefonica I+D - Telefonica Investigación y Desarrollo
4 NWU - North-West University [Vaal Triangle Campus, Vanderbijlpark, South Africa]
5 TEXMEX - Multimedia content-based indexing
6 IIIT-H - International Institute of Information Technology [Hyperabad]
7 METISS - Speech and sound data modeling and processing
8 Speech@FIT
9 HCTLab - Human Computer Technology Laboratory

In this paper, we describe the "Spoken Web Search" Task, which was held as part of the 2011 MediaEval benchmark campaign. The purpose of this task was to perform audio search with audio input in four languages, with very few resources being available in each language. The data was taken from "spoken web" material collected over mobile phone connections by IBM India. We present results from several independent systems, developed by five teams and using different approaches, compare them, and provide analysis and directions for future research.

Keywords: low-resource speech recognition; evaluation; spoken web; audio search; spoken term detection

Domains: Multimedia [cs.MM]

Florian Metze, Nitendra Rajput, Xavier Anguera, Marelie Davel, Guillaume Gravier, et al.. The Spoken Web Search task at MediaEval 2011. ICASSP - 37th International Conference on Acoustics, Speech, and Signal Processing, 2012, France. ⟨hal-00671011⟩