抄録
In this paper, we propose a new image interpolation method based on a 2-D piecewise stationary autoregressive (PAR) model. SAI, which defined PAR model, is a state-of-the-art method in image interpolation. It produces good quality restored images but has high calculation cost because of solving multiple least-square problems. Our method utilizes Gaussian function in estimating parameters instead of solving least-square problems and reduces the calculation cost. Moreover, parameters are estimated at each pixel, while they are estimated in each local window in SAI. By these improvements, the proposed method has equivalent quality to SAI with low calculation cost.
本文言語 | English |
---|---|
ホスト出版物のタイトル | Conference Record - Asilomar Conference on Signals, Systems and Computers |
出版社 | IEEE Computer Society |
ページ | 1193-1197 |
ページ数 | 5 |
巻 | 2016-February |
ISBN(印刷版) | 9781467385763 |
DOI | |
出版ステータス | Published - 2016 2月 26 |
イベント | 49th Asilomar Conference on Signals, Systems and Computers, ACSSC 2015 - Pacific Grove, United States 継続期間: 2015 11月 8 → 2015 11月 11 |
Other
Other | 49th Asilomar Conference on Signals, Systems and Computers, ACSSC 2015 |
---|---|
国/地域 | United States |
City | Pacific Grove |
Period | 15/11/8 → 15/11/11 |
ASJC Scopus subject areas
- コンピュータ ネットワークおよび通信
- 信号処理