Presentation 2021-01-21
[Invited Talk] Lossless Image/Video Coding Method Based on Probability Model Estimation and Optimization
Kyohei Unno,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this talk, the lossless image/video coding method that is proposed by the author is introduced. The proposed method estimates the probability distribution of pixel value by using the Gaussian mixture model for pixel-by-pixel. Then, each pixel value is coded by multiple-value arithmetic coding. Each Gaussian is generated by adaptive prediction and example search. The adaptive prediction is using information neighboring a target pixel. On the other hand, the example search is using non-neighboring information. The shape of the probability distribution can be controlled by a few parameters. To improve coding performance, those parameters are optimized to minimize entropy by a non-linear optimization method.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Lossless coding / Example search / Adaptive prediction / Quasi-Newton method
Paper # IE2020-36
Date of Issue 2021-01-14 (IE)

Conference Information
Committee IE
Conference Date 2021/1/21(1days)
Place (in Japanese) (See Japanese page)
Place (in English) Online
Topics (in Japanese) (See Japanese page)
Topics (in English) Image Processing, Image Coding, etc
Chair Hideaki Kimata(NTT)
Vice Chair Kazuya Kodama(NII) / Keita Takahashi(Nagoya Univ.)
Secretary Kazuya Kodama(KDDI Research) / Keita Takahashi(Nagoya Inst. of Tech.)
Assistant Shunsuke Iwamura(NHK) / Shinobu Kudo(NTT)

Paper Information
Registration To Technical Committee on Image Engineering
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) [Invited Talk] Lossless Image/Video Coding Method Based on Probability Model Estimation and Optimization
Sub Title (in English)
Keyword(1) Lossless coding
Keyword(2) Example search
Keyword(3) Adaptive prediction
Keyword(4) Quasi-Newton method
1st Author's Name Kyohei Unno
1st Author's Affiliation KDDI Research, Inc.(KDDI Research)
Date 2021-01-21
Paper # IE2020-36
Volume (vol) vol.120
Number (no) IE-329
Page pp.pp.8-8(IE),
#Pages 1
Date of Issue 2021-01-14 (IE)