The NNI query-by-Example system for MediaEval 2014

Peng Yang; Haihua Xu; Xiong Xiao; Lei Xie; Cheung Chi Leung; Hongjie Chen; Jia Yu; Hang Lv; Lei Wang; Su Jun Leow; Bin Ma; Eng Siong Chng; Haizhou Li

The NNI query-by-Example system for MediaEval 2014

Peng Yang, Haihua Xu, Xiong Xiao, Lei Xie, Cheung Chi Leung, Hongjie Chen, Jia Yu, Hang Lv, Lei Wang, Su Jun Leow, Bin Ma, Eng Siong Chng, Haizhou Li

School of Computer Science

Research output: Contribution to journal › Conference article › peer-review

10 Scopus citations

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.

Original language	English
Journal	CEUR Workshop Proceedings
Volume	1263
State	Published - 2014
Event	Multimedia Benchmark Workshop, MediaEval 2014 - Barcelona, Spain Duration: 16 Oct 2014 → 17 Oct 2014

Cite this

@article{7504b2b050c34d24a14957d6d91c5498,

title = "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.",

author = "Peng Yang and Haihua Xu and Xiong Xiao and Lei Xie and Leung, {Cheung Chi} and Hongjie Chen and Jia Yu and Hang Lv and Lei Wang and Leow, {Su Jun} and Bin Ma and Chng, {Eng Siong} and Haizhou Li",

year = "2014",

language = "英语",

volume = "1263",

journal = "CEUR Workshop Proceedings",

issn = "1613-0073",

publisher = "CEUR-WS",

note = "Multimedia Benchmark Workshop, MediaEval 2014 ; Conference date: 16-10-2014 Through 17-10-2014",

}

TY - JOUR

T1 - The NNI query-by-Example system for MediaEval 2014

AU - Yang, Peng

AU - Xu, Haihua

AU - Xiao, Xiong

AU - Xie, Lei

AU - Leung, Cheung Chi

AU - Chen, Hongjie

AU - Yu, Jia

AU - Lv, Hang

AU - Wang, Lei

AU - Leow, Su Jun

AU - Ma, Bin

AU - Chng, Eng Siong

AU - Li, Haizhou

PY - 2014

Y1 - 2014

N2 - 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.

AB - 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.

UR - http://www.scopus.com/inward/record.url?scp=84909960178&partnerID=8YFLogxK

M3 - 会议文章

AN - SCOPUS:84909960178

SN - 1613-0073

VL - 1263

JO - CEUR Workshop Proceedings

JF - CEUR Workshop Proceedings

T2 - Multimedia Benchmark Workshop, MediaEval 2014

Y2 - 16 October 2014 through 17 October 2014

ER -

The NNI query-by-Example system for MediaEval 2014

Abstract

Other files and links

Fingerprint

Cite this