39. 宣伝:色線形性に基づく正則化(ADMM)
39
全変動のみ原画像
PSNR: 22.17 PSNR: 26.14
応用:圧縮センシング再構成(2割の情報から復元)
POINS:カラー画像の局所的な色線形性を重み付き核型ノルムでモデル化(非凸)
全変動+提案
S. Ono & I. Yamada “Color-line regularization for color artifact removal,’’ IEEE Trans. Comput. Imag., 2016.
40. 宣伝:グラフ信号用の一般化全変動(PDS)
40
Original
Noisy (sigma = 0.05)
Noisy (sigma = 0.1)
GTV (RMSE = 1.80e-3)
GTV (RMSE = 2.95e-3)
GTGV (RMSE = 1.40e-3)
GTGV (RMSE = 2.20e-3)
応用:歪んだメッシュを平滑化
? グラフ構造は三角メッシュ
? 信号 = xyz座標 (これが歪んでいる)
? グラフ信号用全変動と比較
S. Ono et al., “Total generalized variation for graph signals,” ICASSP 2015.
41. 宣伝:確率的ADMMによる画像復元
41L-ADMMPDS 確率的ADMM (b=32, 64, 128)
? b : 観測行列Φの分割数
? NRMSEn := ||u(n) – u*||/||u*||
(最適解との正規化誤差)
提案法が10倍近く高速
CPU time
NRMSEn
PDS
LADMM
Prop (b=32)
Prop (b=64)
Prop (b=128)
S. Ono et al., “Image restoration using a
stochastic variant of the alternating
direction method of multipliers,” ICASSP 2016.
復元画像の品質は非確率
的最適化の場合と同等
応用:観測行列Φが密な場合の効率的画像復元
43. 参考文献 (1/4)
43
◆近接勾配法 (proximal gradient, forward-backward splitting)
1. G. B. Passty, “Ergodic convergence to a zero of the sum of monotone operators
in Hilbert space,” J. Math. Anal. Appl., 1979. (原著)
2. G. Chen & R. T. Rockafellar, “Convergence rates in forward–backward splitting,”
SIAM J. Optim., 1997. (収束レート解析)
3. P. L. Combettes & V. R. Wajs, ”Signal recovery by proximal forward–backward
splitting,” SIAM Multiscale Model. Simul., 2005. (更なる理論的整備)
4. A. Beck & M. Teboulle, “A fast iterative shrinkage-thresholding algorithm for
linear inverse problems,” SIAM J. Imag. Sci., 2009. (通称 FISTA, Nesterov の optimal
gradient を融合)
5. M. Yamagishi & I. Yamada, “Over-relaxation of the fast iterative shrinkage-
thresholding algorithm with variable stepsize,” Inverse Probl., 2011. (FISTA のス
テップサイズ拡張)
6. I. Daubechies et al., “An iterative thresholding algorithm for linear inverse
problems with a sparsity constraint,” Comm. Pure Appl. Math., 2004. (スパース復
元応用)
7. J. Duchi & Y. Singer, “Efficient online and batch learning using forward-
backward splitting,” J Mach. Learn. Res., 2009. (機械学習応用)
44. 参考文献 (2/4)
44
◆ADMM (Alternating Direction Method of Multipliers)
1. D. Gabay & B. Mercier, “A dual algorithm for the solution of nonlinear
variational problems via finite elements approximations,” Comput. Math. Appl.,
1976. (原著)
2. J. Eckstein & D. P. Bertsekas, “On the Douglas-Rachford splitting method and the
proximal point algorithm for maximal monotone operators,” Math. Program.,
1992. (Dougal-Rachford splittingに基づく導出と理論的整備)
3. B. He & X. Yuan “On the O(1/n) Convergence Rate of the Douglas–Rachford
Alternating Direction Method,” SIAM J. Numer. Anal., 2012. (収束レート)
4. S. Boyd et al., “Distributed optimization and statistical learning via the
alternating direction method of multipliers,” Found. Trends Mach. Learn., 2011.
(分散最適化?機械学習適用とレビュー)
5. M. Afonso et al. “An augmented Lagrangian approach to the constrained
optimization formulation of imaging inverse problems,” IEEE Trans. Image
Process., 2011. (画像復元応用)
6. S. Ono et al., “Cartoon-texture image decomposition using blockwise low-rank
texture characterization,” IEEE Trans. Image Process., 2014. (画像分離応用)
7. S. Ono et al., “Image restoration using a stochastic variant of the alternating
direction method of multipliers,” ICASSP 2016. (確率的ADMMの画像復元応用)
8. S. Ono & I. Yamada “Color-line regularization for color artifact removal,’’ IEEE
Trans. Comput. Imag., 2016. (色線形性正則化,非凸応用)
45. 参考文献 (3/4)
45
◆主-双対近接分離法 (primal-dual splitting, PDS)
1. A. Chambolle & T. Pock, “A first-order primal-dual algorithm for convex
problems with applications to imaging,” J. Math. Imag. Vis., 2011. (原著、少し限
定的な形)
2. L. Condat, “A primal-dual splitting method for convex optimization involving
Lipschitzian, proximable and linear composite terms,” J. Optim. Theory Appl.,
2013. (原著、今回紹介した形)
3. R. Bo? & E. Csetnek, “On the convergence rate of a forward-backward type
primal-dual splitting algorithm for convex optimization problems,”
Optimization, 2015. (収束レート)
4. S. Ono and I. Yamada, “A convex regularizer for reducing color artifact in color
image recovery,” CVPR 2013. (色ムラ?偽色除去)
5. S. Ono and I. Yamada, “Decorrelated vectorial total variation,” CVPR, 2014. (カ
ラー画像処理応用)
6. S. Ono, M. Yamagishi, and I. Yamada, “A sparse system identification by using
adaptively-weighted total variation via a primal-dual splitting approach,”
ICASSP 2013. (adaptive版)
7. S. Ono, I. Yamada, and I. Kumazawa, “Total generalized variation for graph signals,”
ICASSP 2015. (グラフ信号処理応用)
8. S. Ono and I. Yamada, “Hierarchical convex optimization with primal-dual
splitting,” IEEE Trans. Signal Process., 2015. (PDSを利用した階層型最適化)
46. 参考文献 (4/4)
46
◆その他、関係する and/or 役立つと思われる文献?書籍
1. J. J. Moreau, “Fonctions convexes duales et points proximaux dans un espace
Hilbertien,” (in French) C. R. Acad. Sci. Paris Ser. AMath., 1962. (近接写像初出)
2. H. H. Bauschke & P. L. Combettes, Convex analysis and monotone operator
theory in Hilbert spaces. Springer, 2011. (凸解析と単調写像理論の関係について
体系的にまとめた良書、議論は基本的に無限次元ヒルベルト空間で展開)
3. P. L. Combettes & J.-C. Pesquet, “Proximal splitting methods in signal
processing,” in Fixed-Point Algorithm for Inverse Problems in Science and
Engineering, Springer-Verlag, 2011. (近接分離最適化のレビュー、近接写像が効率
的に計算可能な凸関数のリスト)
4. I. Yamada et al., “Minimizing the Moreau envelope of nonsmooth convex
functions over the fixed point set of certain quasi-nonexpansive mappings,” in
Fixed-Point Algorithm for Inverse Problems in Science and Engineering, Springer-
Verlag, 2011. (いくつかの近接分離系の手法を非拡大写像の不動点的視点から解説)
5. N. Parikh & S. Boyd, “Proximal algorithms,“ Foundations and Trends in
Optimization, 2014. (近接分離最適化のレビュー)
6. S. Ono & I. Yamada, “Signal recovery with certain involved convex data-fidelity
constraints,” IEEE Trans. Signal Process., 2015. (近接分離系の手法では扱えないデー
タ忠実性制約が課された信号復元問題を解く最適化アルゴリズムの提案)