Presentation 2016-06-16
Image retrieval in JPEG 2000 codestream domain by a deep learning approach
Yusuke Sugawara, Osamu Watanabe, Sayaka Shiota, Hitoshi Kiya,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) A codestream-level image retrieval scheme for JPEG 2000 images byusing a deep learning approach is proposed in thisarticle. The aim is to retrieve similar images to a query imagecompressed with the JPEG 2000 standard, from a database in which JPEG2000 images are labeled according to the similarity among images. Theproposed retrieval scheme can be performed without decoding images because the features used in theimage retrieval can be directly extractedfrom JPEG 2000 codestream. In previous research, a codestram-levelretrieval scheme has been studied for JPEG 2000 images. However, itdoes not have enough retrieval accuracy, when there are no queryimages in the database. The proposed method can improve this problem. A number of simulations are carried out to confirm the effectivenessof the proposed scheme.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) JPEG 2000 / Deep Learning / Image Retrieval
Paper # CAS2016-15,VLD2016-21,SIP2016-49,MSS2016-15
Date of Issue 2016-06-09 (CAS, VLD, SIP, MSS)

Conference Information
Committee VLD / CAS / MSS / SIP
Conference Date 2016/6/16(2days)
Place (in Japanese) (See Japanese page)
Place (in English) Hirosaki Shiritsu Kanko-kan
Topics (in Japanese) (See Japanese page)
Topics (in English) System, signal processing and related topics
Chair Takashi Takenana(NEC) / Toshihiko Takahashi(Niigata Univ.) / Satoshi Yamane(Kanazawa Univ.) / Makoto Nakashizuka(Chiba Inst. of Tech.)
Vice Chair Hiroyuki Ochi(Ritsumeikan Univ.) / Mitsuru Hiraki(Renesas) / Morikazu Nakamura(Univ. of Ryukyus) / Masahiro Okuda(Univ. of Kitakyushu) / Shogo Muramatsu(Niigata Univ.)
Secretary Hiroyuki Ochi(Fujitsu Labs.) / Mitsuru Hiraki(Hiroshima City Univ.) / Morikazu Nakamura(Tohoku Univ.) / Masahiro Okuda(Renesas) / Shogo Muramatsu(Yamaguchi Univ.)
Assistant Parizy Matthieu(Fujitsu Labs.) / Toshihiro Tachibana(Shonan Inst. of Tech.) / Yohei Nakamura(Hitachi) / Hideki Kinjo(Okinawa Univ.) / Osamu Watanabe(Takushoku Univ.)

Paper Information
Registration To Technical Committee on VLSI Design Technologies / Technical Committee on Circuits and Systems / Technical Committee on Mathematical Systems Science and its applications / Technical Committee on Signal Processing
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Image retrieval in JPEG 2000 codestream domain by a deep learning approach
Sub Title (in English)
Keyword(1) JPEG 2000
Keyword(2) Deep Learning
Keyword(3) Image Retrieval
1st Author's Name Yusuke Sugawara
1st Author's Affiliation Tokyo Metropolitan University(Tokyo Metropolitan Univ)
2nd Author's Name Osamu Watanabe
2nd Author's Affiliation Takushoku University(Takushoku Univ)
3rd Author's Name Sayaka Shiota
3rd Author's Affiliation Tokyo Metropolitan University(Tokyo Metropolitan Univ)
4th Author's Name Hitoshi Kiya
4th Author's Affiliation Tokyo Metropolitan University(Tokyo Metropolitan Univ)
Date 2016-06-16
Paper # CAS2016-15,VLD2016-21,SIP2016-49,MSS2016-15
Volume (vol) vol.116
Number (no) CAS-93,VLD-94,SIP-95,MSS-96
Page pp.pp.79-84(CAS), pp.79-84(VLD), pp.79-84(SIP), pp.79-84(MSS),
#Pages 6
Date of Issue 2016-06-09 (CAS, VLD, SIP, MSS)