Presentation | 2018-09-20 Image-Caption Retrieval by Embedding to Gaussian Distribution Kenta Hama, Takashi Matsubara, Kuniaki Uehara, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | To get distributed representations of words, one has typically embedded words to points. Recent studies successfully represent the uncertainty of the meaning of words by embedding them to probability distributions. We consider such representations contribute to image-caption retrieval. With point embedding, a ``bus'' is considered to be similar to but different from the ``vehicle''. Since the distribution embedding captures the ambiguous meaning of the word ``vehicle'', a ``bus'' is included in the``vehicle'', and one can get a ``bus'' image using a text query of ``vehicle''. In this study, we propose a method to embed image and text data to a normal distribution. Compared to point embedding, our proposed embedding improves the performances on the image-caption retrieval benchmark tasks. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Representation Learning / Image-Caption Retrieval / Embeddings |
Paper # | PRMU2018-38,IBISML2018-15 |
Date of Issue | 2018-09-13 (PRMU, IBISML) |
Conference Information | |
Committee | PRMU / IBISML / IPSJ-CVIM |
---|---|
Conference Date | 2018/9/20(2days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | Shinichi Sato(NII) / Hisashi Kashima(Kyoto Univ.) |
Vice Chair | Yoshihisa Ijiri(Omron) / Toru Tamaki(Hiroshima Univ.) / Masashi Sugiyama(Univ. of Tokyo) / Koji Tsuda(Univ. of Tokyo) |
Secretary | Yoshihisa Ijiri(NEC) / Toru Tamaki(Osaka Univ.) / Masashi Sugiyama(Nagoya Inst. of Tech.) / Koji Tsuda(AIST) |
Assistant | Go Irie(NTT) / Yoshitaka Ushiku(Univ. of Tokyo) / Tomoharu Iwata(NTT) / Shigeyuki Oba(Kyoto Univ.) |
Paper Information | |
Registration To | Technical Committee on Pattern Recognition and Media Understanding / Technical Committee on Infomation-Based Induction Sciences and Machine Learning / Special Interest Group on Computer Vision and Image Media |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Image-Caption Retrieval by Embedding to Gaussian Distribution |
Sub Title (in English) | |
Keyword(1) | Representation Learning |
Keyword(2) | Image-Caption Retrieval |
Keyword(3) | Embeddings |
1st Author's Name | Kenta Hama |
1st Author's Affiliation | Kobe University(Kobe Univ.) |
2nd Author's Name | Takashi Matsubara |
2nd Author's Affiliation | Kobe University(Kobe Univ.) |
3rd Author's Name | Kuniaki Uehara |
3rd Author's Affiliation | Kobe University(Kobe Univ.) |
Date | 2018-09-20 |
Paper # | PRMU2018-38,IBISML2018-15 |
Volume (vol) | vol.118 |
Number (no) | PRMU-219,IBISML-220 |
Page | pp.pp.17-20(PRMU), pp.17-20(IBISML), |
#Pages | 4 |
Date of Issue | 2018-09-13 (PRMU, IBISML) |