The NNI Query-by-Example System for MediaEval 2014

Abstract

In this paper we describe the system proposed by NNI (NWPU-NTU-I2R) team for the QUESST task within the Mediaeval 2014 evaluation. To solve the problem, we used both dy-namic time warping (DTW) and symbolic search (SS) based approaches. The DTW system performs template match-ing using subsequence DTW algorithm and posterior repre-sentations. The symbolic search is performed on phone se-quences generated by phone recognizers. For both symbolic and DTW search, partial sequence matching is performed to reduce missing rate, especially for query type 2 and 3. Af-ter fusing 9 DTW systems, 7 symbolic systems, and query length side information, we obtained 0.6023 actual normal-ized cross entropy (actCnxe) for all queries combined. For type 3 complex queries, we achieved 0.7252 actCnxe. 1.