Presentation	2022-03-10 [Special Talk] Lossless Image Coding using Inpainting-Oriented Deep Pixel Predictor Keita Takahashi,
PDF Download Page	PDF download Page Link
Abstract(in Japanese)	(See Japanese page)
Abstract(in English)	I will be presenting our previous paper that received IE special Award 2020 to encourage discussions for future directions. The coding efficiency of lossless image coding methods depends on the pixel predictor that is designed to predict probabilistic distributions for unknown pixels from the already-processed pixels. Recently, more attention is given to deep pixel predictors, which are optimized using abundant training data under the framework of deep learning. To achieve high coding efficiency and fast computation speed, we focus on the processing order of pixels in deep pixel predictors. In previous works, the image pixels are usually processed in the raster-scan order starting from the top-left pixel. This processing order restricts the reference area used to predict a target pixel only to the upper part of that pixel. Moreover, due to this processing order, the pixels need to be predicted and decoded one by one at the decoding time, which incurs significant computation time. Meanwhile, we construct a deep pixel predictor that can predict multiple target pixels simultaneously by referring to the pixels located at all the directions with respect to the target pixels. Our pixel predictor is designed as a stack of deep CNNs that were used for image inpainting task, in which the number of the processed pixels gradually increases as the image undergoes more CNNs. Our experimental results show that our inpainting-oriented predictor can achieve the coding efficiency comparable to that of the raster-scan order counterpart and substantially decrease the computation time required for the decoding process.
Keyword(in Japanese)	(See Japanese page)
Keyword(in English)	Lossless image coding / Deep pixel predictor / image inpainting
Paper #	IMQ2021-31,CQ2021-122,IE2021-93,MVE2021-60
Date of Issue	2022-03-02 (IMQ, CQ, IE, MVE)

Conference Information
Committee	CQ / IMQ / MVE / IE
Conference Date	2022/3/9(3days)
Place (in Japanese)	(See Japanese page)
Place (in English)	Online (Zoom)
Topics (in Japanese)	(See Japanese page)
Topics (in English)	Media of five senses, Multimedia, Media experience, Picture codinge, Image media quality, Network,quality and reliability, etc
Chair	Jun Okamoto(NTT) / Kenya Uomori(Osaka Univ.) / Masayuki Ihara(RIKEN) / Kazuya Kodama(NII)
Vice Chair	Takefumi Hiraguri(Nippon Inst. of Tech.) / Gou Hasegawa(Tohoku Univ.) / Mitsuru Maeda(Canon) / Kiyoshi Kiyokawa(NAIST) / Hiroyuki Bandoh(NTT) / Toshihiko Yamazaki(Univ. of Tokyo)
Secretary	Takefumi Hiraguri(NTT) / Gou Hasegawa(Ritsumeikan Univ.) / Mitsuru Maeda(Nagoya Univ.) / Kiyoshi Kiyokawa(NTT) / Hiroyuki Bandoh(Oosaka Inst. of Tech.) / Toshihiko Yamazaki(NTT)
Assistant	Yoshiaki Nishikawa(NEC) / Ryoichi Kataoka(KDDI Research) / Kimiko Kawashima(NTT) / Masato Tsukada(NEC) / Takashi Yamazoe(Seikei Univ.) / Naoya Isoyama(NAIST) / Takenori Hara(DNP) / Mitsuhiro Goto(NTT) / Shunsuke Iwamura(NHK) / Shinobu Kudo(NTT)

Paper Information
Registration To	Technical Committee on Communication Quality / Technical Committee on Image Media Quality / Technical Committee on Media Experience and Virtual Environment / Technical Committee on Image Engineering
Language	JPN
Title (in Japanese)	(See Japanese page)
Sub Title (in Japanese)	(See Japanese page)
Title (in English)	[Special Talk] Lossless Image Coding using Inpainting-Oriented Deep Pixel Predictor
Sub Title (in English)
Keyword(1)	Lossless image coding
Keyword(2)	Deep pixel predictor
Keyword(3)	image inpainting
1st Author's Name	Keita Takahashi
1st Author's Affiliation	Nagoya University(Nagoya Univ.)
Date	2022-03-10
Paper #	IMQ2021-31,CQ2021-122,IE2021-93,MVE2021-60
Volume (vol)	vol.121
Number (no)	IMQ-420,CQ-421,IE-422,MVE-423
Page	pp.pp.114-114(IMQ), pp.124-124(CQ), pp.114-114(IE), pp.114-114(MVE),
#Pages	1
Date of Issue	2022-03-02 (IMQ, CQ, IE, MVE)