上げ直します

1. 2016/12/03 CV勉強会@関東
ECCV2016読み会 発表資料
2016/12/03
@peisuke

2. 自己紹介
名前：藤本 敬介
研究：ロボティクス、コンピュータビジョン
点群：形状統合、メッシュ化、認識
画像：画像認識、SfM・MVS
ロボット：自律移動、動作計画

3. 本発表の概要
• 発表論文
• Sublabel-Accurate Convex Relaxation of Vectorial
Multilabel Energies
• どんな論文？
• 多次元のラベリング問題を凸緩和によって効率良く解く
• ノイズ除去、密オプティカルフロー、ステレオマッチング等
• 特徴は？
• 問題を複数の区間に分け、区間ごとに凸関数で近似
• 一般的な式に対する解法なので汎用性高
※資料が間に合わなかった
ので、後で挙げ直します

4. Sublabel-Accurate Convex
Relaxation of Vectorial Multilabel
Energies
Emanuel Laude, Thomas Möllenhoff, Michael
Moeller, Jan Lellmann, Daniel Cremers

5. 目的とする問題の解説
• どんな問題に使える？
• 画像の各画素𝑥に対して、どのラベル𝑢 𝑥 が正しいか
について、何らかの評価関数𝜌が与えられているとする
• 評価関数の値を良くしつつ、隣り合う画素同士のラベル
は近いようにする
𝐸 = min
):+→-
. 𝜌 𝑥, 𝑢 𝑥 𝑑𝑥 + 𝜆Ψ 𝛻𝑢
+
Ψ 𝛻𝑢 = 𝑇𝑉(𝑢) = ∫ 𝛻𝑢 𝑥+ ;
𝑑𝑥
近い値ほど良い関数何らかの評価関数
例えば・・・

6. 目的とする問題の解説
• どんな問題に使える？
• 画像の各画素𝑥に対して、どのラベル𝑢 𝑥 が正しいか
について、何らかの評価関数𝜌が与えられているとする
• 評価関数の値を良くしつつ、隣り合う画素同士のラベル
は近いようにする
𝐸 = min
):+→-
. 𝜌 𝑥, 𝑢 𝑥 𝑑𝑥 + 𝜆Ψ 𝛻𝑢
+
Ψ 𝛻𝑢 = 𝑇𝑉(𝑢) = ∫ 𝛻𝑢 𝑥+ ;
𝑑𝑥
近い値ほど良い関数何らかの評価関数
例えば・・・

7. 事例：ステレオマッチング
この点と同じ模様をこのライン上から探し、視差を計算
「この点」上に、画素の一致度に関し、こんな評価関数が定義できる

8. 事例：ステレオマッチング
全ての画素に対して、それぞれ評価関数を算出

9. 事例：ステレオマッチング
隣り合う画素同士の視差の変化が小さくなるようにしつつ、
評価の高い所を選択

10. 事例：オプティカルフロー
• 動画像中の各画素が、次の時点でどの方向にど
れだけ動くかを、各画素とマッチする周辺位置を探
索することで求める
xy座標に対するマッチ度合い

11. 事例：オプティカルフロー
• ステレオマッチングと同様に隣り合う画素同士のフ
ローの変化が小さくなるようにしつつ、マッチ度を
高くする

12. 目的とする問題の解説
• どんな問題に使える？
• 画像の各画素𝑥に対して、どのラベル𝑢 𝑥 が正しいか
について、何らかの評価関数𝜌が与えられているとする
• 評価関数の値を良くしつつ、隣り合う画素同士のラベル
は近いようにする
𝐸 = min
):+→-
. 𝜌 𝑥, 𝑢 𝑥 𝑑𝑥 + 𝜆Ψ 𝛻𝑢
+
Ψ 𝛻𝑢 = 𝑇𝑉(𝑢) = ∫ 𝛻𝑢 𝑥+ ;
𝑑𝑥
近い値ほど良い関数何らかの評価関数
例えば・・・

13. エッジを残す平滑化（TVノルム）
• 隣接画素間の２乗距離を使うとエッジがぼやけて
しまうが、絶対値を使うとエッジが保存される
𝐸 𝐮 = ℓ 𝐮 + 𝜆Ω 𝛁𝐮 Ω 𝛁𝐮 = @ 𝛻𝑢A + 𝛻𝑢B
uu
0
1
0
1
どちらも|∇u|=１
0
1
元の信号と近く、
制約項のペナルティも小
元信号
ノイズ除去後
信号候補１
ノイズ除去後
信号候補２

14. （参考）Vectorial Total Variation
• 多チャンネル画像では、チャンネル毎に異なる向
きのエッジが出力され、ボヤけてしまう
• チャンネル間でエッジは共通して残しつつ、変化量
のみチャンネル毎に計算する事のできる評価関数
• ∇uの最大の特異値をノルムとする

15. 目的とする問題（再掲）
• 多値ラベリング問題・・・直接解くのは困難
• 高次元・非線形・微分不可能
𝐸 = min
):+→-
. 𝜌 𝑥, 𝑢 𝑥 𝑑𝑥 + 𝜆Ψ 𝛻𝑢
+
解きやすい凸問題で近似しよう

16. 凸問題とは
• 局所最適解＝大局最適解となるような単純な問題
• 最適化が容易
非凸問題 凸問題

17. 近似方法について
• 目的関数fの双共役（双対の双対）を取ると、fを
下から抑える凸関数f**に変換できる
𝑓 𝒙 𝑓∗∗
𝒙

18. 双対とは
• 関数𝑓∗
は𝑓の共役関数と呼ばれる
• 𝑓∗
の最大化は𝑓の最小化と等価
𝑓∗
𝒔 = sup 𝒔K
𝒙 − 𝑓 𝒙 | 𝒙 ∈ 𝑹P

19. Keywords: Convex Relaxation, Optim
(a) Original dataterm (b) Without lifting (
arXiv:16
双共役による近似式
• このまま双共役にすると近似精度低
𝐸 = min
):+→-
. 𝜌∗∗ 𝑥, 𝑢 𝑥 𝑑𝑥 + 𝜆Ψ∗∗ 𝛻𝑢
+
𝜌∗∗
𝑥, 𝑢 𝑥𝜌 𝑥, 𝑢 𝑥

20. (c) Classical lifting (d) Proposed lifting
ataterm. Convexification without lifting
領域分けによる高精度な近似
近似精度は高くなるが全体としては非凸関数
→問題の次元数を増やし凸問題のまま上記を解く
領域を３角形の集合で分割
（たぶん・・・！）

21. Lifted Representation
• 変数を高次元の変数に変換
• 変数uを、uの属する３角形の頂点の線形和で表す
• u = (0.3, 0.2) → 0.7 * t2 + 0.1 * t3 + 0.2 * t6
• 使わなかった頂点はゼロとして、要素数が頂点数のベ
クトルに変換
• (0, 0.7, 0.1, 0, 0, 0.2)
Sublabel-Accurate Convex Relaxation of Vectorial M
1 0 1
0
1
u(x) = 0.7e
2
+ 0.1e
3
+ 0.2e
6
= (0, 0.7, 0.1, 0, 0, 0.2)
>
1
4
0.2
0.3
t6
t2
t3
𝑢 = @ 𝒖R 𝑡R

22. Lifted Representationでの問題設定
• 高次元化されたデータ項と正則化項
𝝆 𝒖 = U
𝜌 𝑇R 𝛼 ,
∞,
Ψ 𝑔 = U 𝑇R 𝛼 − 𝑇Y 𝛽 [ 𝜐
∞
𝑖𝑓 𝒖 = 𝐸R 𝛼
𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
𝑖𝑓 𝑔 = 𝐸R 𝛼 − 𝐸Y 𝛽 ⊗ 𝜐
𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
. 𝝆∗∗
𝑥, 𝒖 𝑥 + 𝜳∗∗
𝛻𝒖 𝑑𝑥

23. Lifted Representationでの問題設定
• 高次元化されたデータ項と正則化項
𝝆 𝒖 = U
𝜌 𝑇R 𝛼 ,
∞,
Ψ 𝑔 = U 𝑇R 𝛼 − 𝑇Y 𝛽 [ 𝜐
∞
𝑖𝑓 𝒖 = 𝐸R 𝛼
𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
𝑖𝑓 𝑔 = 𝐸R 𝛼 − 𝐸Y 𝛽 ⊗ 𝜐
𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
. 𝝆∗∗
𝑥, 𝒖 𝑥 + 𝜳∗∗
𝛻𝒖 𝑑𝑥

24. Lifted Representationの具体例
𝝆 𝒖 = U
𝜌 𝑇R 𝛼 ,
∞,
𝑖𝑓 𝒖 = 𝐸R 𝛼
𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
𝒖 = 0, 0.7, 0.1,0, 0, 0.2 の場合・・・
𝐸; = 𝑒;, 𝑒k, 𝑒l
𝐸m = 𝑒;, 𝑒m, 𝑒l
𝐸n = 𝑒m, 𝑒l, 𝑒o
𝐸k = 𝑒m, 𝑒n, 𝑒o
t1=(0,0) t2=(1,0) t3=(2,0)
t4=(0,1) t5=(1,1) t6=(2,1)
∆;
∆m
∆n
∆k
𝒖 = 0.7𝑒m + 0.1𝑒n + 0.2𝑒o
𝑒m = 0, 1, 0, 0, 0, 0
元空間での表現
Liftedされた表現
𝛼 = 0.7, 0.1, 0.2
𝝆 𝒖 = 𝜌 𝑢 ,
𝑢 = 0.7𝑡m + 0.1𝑡n + 0.2𝑡o

25. Lifted Representationの具体例
𝝆 𝒖 = U
𝜌 𝑇R 𝛼 ,
∞,
𝑖𝑓 𝒖 = 𝐸R 𝛼
𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
𝒖 = 0.5, 0.7, 0.1,0.2, 0.1, 0.2 の場合・・・
𝐸; = 𝑒;, 𝑒k, 𝑒l
𝐸m = 𝑒;, 𝑒m, 𝑒l
𝐸n = 𝑒m, 𝑒l, 𝑒o
𝐸k = 𝑒m, 𝑒n, 𝑒o
t1=(0,0) t2=(1,0) t3=(2,0)
t4=(0,1) t5=(1,1) t6=(2,1)
∆;
∆m
∆n
∆k
𝑒m = 0, 1, 0, 0, 0, 0
元空間での表現
Liftedされた表現
𝝆 𝒖 = ∞
無い！

26. Lifted Representationでの問題設定
• 高次元化されたデータ項と正則化項
𝝆 𝒖 = U
𝜌 𝑇R 𝛼 ,
∞,
Ψ 𝑔 = U 𝑇R 𝛼 − 𝑇Y 𝛽 [ 𝜐
∞
𝑖𝑓 𝒖 = 𝐸R 𝛼
𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
𝑖𝑓 𝑔 = 𝐸R 𝛼 − 𝐸Y 𝛽 ⊗ 𝜐
𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
. 𝝆∗∗
𝑥, 𝒖 𝑥 + 𝜳∗∗
𝛻𝒖 𝑑𝑥

27. Lifted Representationの具体例
t1=(0,0) t2=(1,0) t3=(2,0)
t4=(0,1) t5=(1,1) t6=(2,1)
∆;
∆k
Ψ 𝑔 = U 𝑇R 𝛼 − 𝑇Y 𝛽 [ 𝜐
∞
𝑖𝑓 𝑔 = 𝐸R 𝛼 − 𝐸Y 𝛽 ⊗ 𝜐
𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
𝑇R 𝛼
𝑇Y 𝛽 𝜐
画像上のエッジ
※𝜐は正則化項の双対に出てくる補助変数と等価、計算上は陽に出て来ない
[ ; = sup @ 𝑢Rdiv
R
𝜂R 𝝊 − 𝛿
𝛻𝑢

28. Liftedした評価関数について
. 𝝆∗∗
𝑥, 𝒖 𝑥 + 𝜳∗∗
𝛻𝒖 𝑑𝑥

29. Liftedした評価関数について
. 𝝆∗∗
𝑥, 𝒖 𝑥 + 𝜳∗∗
𝛻𝒖 𝑑𝑥
axed energy minimization problem becomes
min
u:⌦!R|V|
max
q:⌦!K
X
x2⌦
⇢⇤⇤
(x, u(x)) + hDiv q, ui.
order to get rid of the pointwise maximum over ⇢⇤
i (v) in Eq. (8), we intr
ditional variables w(x) 2 R and additional constraints (v(x), w(x)) 2 C,
that w(x) attains the value of the pointwise maximum:
min
u:⌦!R|V|
max
(v,w):⌦!C
q:⌦!K
X
x2⌦
hu(x), v(x)i w(x) + hDiv q, ui,
ere the set C is given as
C =
1i|T |
Ci, Ci :=
n
(x, y) 2 R|V|+1
| ⇢⇤
i (x)  y
o
.
numerical optimization we use a GPU-based implementation1
of a first
mal-dual method [14]. The algorithm requires the orthogonal projectio
dual variables onto the sets C respectively K in every iteration. Howeve
elaxed energy minimization problem becomes
min
u:⌦!R|V|
max
q:⌦!K
X
x2⌦
⇢⇤⇤
(x, u(x)) + hDiv q, ui. (18)
n order to get rid of the pointwise maximum over ⇢⇤
i (v) in Eq. (8), we introduce
dditional variables w(x) 2 R and additional constraints (v(x), w(x)) 2 C, x 2 ⌦
o that w(x) attains the value of the pointwise maximum:
min
u:⌦!R|V|
max
(v,w):⌦!C
q:⌦!K
X
x2⌦
hu(x), v(x)i w(x) + hDiv q, ui, (19)
where the set C is given as
C =
1i|T |
Ci, Ci :=
n
(x, y) 2 R|V|+1
| ⇢⇤
i (x)  y
o
. (20)
For numerical optimization we use a GPU-based implementation1
of a first-order
rimal-dual method [14]. The algorithm requires the orthogonal projections of
he dual variables onto the sets C respectively K in every iteration. However, the
rojection onto an epigraph of dimension |V| + 1 is di cult for large values of
V|. We rewrite the constraints (v(x), w(x)) 2 Ci, 1  i  |T |, x 2 ⌦ as (n + 1)-
imensional epigraph constraints introducing variables ri
(x) 2 Rn
, si(x) 2 R:
8 E. Laude, T. M¨ollenho↵, M. Moeller, J. Lellmann, D. Creme
Proof. Follows from a calculation starting at the definition of the convex conju-
gate ⇤
. See Appendix A.
Interestingly, although in its original formulation (14) the set K has infinitely
many constraints, one can equivalently represent K by finitely many.
Proposition 3 The set K in equation (14) is the same as
K =
n
q 2 Rd⇥|V|
| Di
q S1  1, 1  i  |T |
o
, Di
q = QiD (TiD) 1
, (15)
where the matrices QiD 2 Rd⇥n
and TiD 2 Rn⇥n
are given as
QiD := qi1
qin+1
, . . . , qin
qin+1
, TiD := ti1
tin+1
, . . . , tin
tin+1
.
Proof. Similar to the analysis in [11], equation (14) basically states the Lipschitz
8 E. Laude, T. M¨ollenho↵, M. Moeller, J. Lellmann, D. Cremers
Proof. Follows from a calculation starting at the definition of the convex conju-
gate ⇤
. See Appendix A.
Interestingly, although in its original formulation (14) the set K has infinitely
many constraints, one can equivalently represent K by finitely many.
Proposition 3 The set K in equation (14) is the same as
K =
n
q 2 Rd⇥|V|
| Di
q S1  1, 1  i  |T |
o
, Di
q = QiD (TiD) 1
, (15)
where the matrices QiD 2 Rd⇥n
and TiD 2 Rn⇥n
are given as
QiD := qi1
qin+1
, . . . , qin
qin+1
, TiD := ti1
tin+1
, . . . , tin
tin+1
.
et for now the weight of the regularizer in (1) be zero. Then, at each point
2 ⌦ we minimize a generally nonconvex energy over a compact set ⇢ Rn
:
min
u2
⇢(u). (6)
We set up the lifted energy so that it attains finite values if and only if the
rgument u is a sparse representation u = Ei↵ of a sublabel u 2 :
⇢(u) = min
1i|T |
⇢i(u), ⇢i(u) =
8
<
:
⇢(Ti↵), if u = Ei↵, ↵ 2 U
n ,
1, otherwise.
(7)
roblems (6) and (7) are equivalent due to the one-to-one correspondence of
= Ti↵ and u = Ei↵. However, energy (7) is finite on a nonconvex set only. In
rder to make optimization tractable, we minimize its convex envelope.
Proposition 1 The convex envelope of (7) is given as:
⇢⇤⇤
(u) = sup
v2R|V|
hu, vi max
1i|T |
⇢⇤
i (v),
⇢⇤
i (v) = hEibi, vi + ⇢⇤
i (A>
i E>
i v), ⇢i := ⇢ + i
.
(8)
and Ai are given as bi := Mn+1
i , Ai := M1
i , M2
i , . . . , Mn
i , where Mj
i are
he columns of the matrix Mi := (T>
i , 1) >
2 Rn+1⇥n+1
.
roof. Follows from a calculation starting at the definition of ⇢⇤⇤
. See Ap-
min
u2
⇢(u). (6)
p the lifted energy so that it attains finite values if and only if the
u is a sparse representation u = Ei↵ of a sublabel u 2 :
= min
1i|T |
⇢i(u), ⇢i(u) =
8
<
:
⇢(Ti↵), if u = Ei↵, ↵ 2 U
n ,
1, otherwise.
(7)
(6) and (7) are equivalent due to the one-to-one correspondence of
nd u = Ei↵. However, energy (7) is finite on a nonconvex set only. In
make optimization tractable, we minimize its convex envelope.
ion 1 The convex envelope of (7) is given as:
⇢⇤⇤
(u) = sup
v2R|V|
hu, vi max
1i|T |
⇢⇤
i (v),
⇢⇤
i (v) = hEibi, vi + ⇢⇤
i (A>
i E>
i v), ⇢i := ⇢ + i
.
(8)
are given as bi := Mn+1
i , Ai := M1
i , M2
i , . . . , Mn
i , where Mj
i are
ns of the matrix Mi := (T>
i , 1) >
2 Rn+1⇥n+1
.
⇤⇤
We set up the lifted energy so that it attains fi
argument u is a sparse representation u = Ei↵ o
⇢(u) = min
1i|T |
⇢i(u), ⇢i(u) =
8
<
:
⇢(Ti↵),
1,
Problems (6) and (7) are equivalent due to the
u = Ti↵ and u = Ei↵. However, energy (7) is fin
order to make optimization tractable, we minimi
Proposition 1 The convex envelope of (7) is gi
⇢⇤⇤
(u) = sup
v2R|V|
hu, vi max
1i|T |
⇢⇤
i
⇢⇤
i (v) = hEibi, vi + ⇢⇤
i (A>
i E>
i v),
bi and Ai are given as bi := Mn+1
i , Ai := M1
i ,
the columns of the matrix Mi := (T>
i , 1) >
2 Rn
Proof. Follows from a calculation starting at t

30. 一見難しそうに見えるけど・・・

31. 一見難しそうに見えるけど・・・
むずかしい！！

32. 凸問題を解く際の戦略
• 主問題の変数で最適化と双対問題の変数で最適
化を交互に繰り返す
• 主問題の変数とは、元々の問題にあった変数
• 双対問題の変数は、後から加えた変数
𝑓∗
𝒔 = sup 𝒔K
𝒙 − 𝑓 𝒙 | 𝒙 ∈ 𝑹P𝑓 𝒙
双対
sが双対問題の変数

33. 今回の問題について
rder to get rid of the pointwise maximum over ⇢⇤
i (v) in Eq. (8), we intro
tional variables w(x) 2 R and additional constraints (v(x), w(x)) 2 C, x
hat w(x) attains the value of the pointwise maximum:
min
u:⌦!R|V|
max
(v,w):⌦!C
q:⌦!K
X
x2⌦
hu(x), v(x)i w(x) + hDiv q, ui,
re the set C is given as
C =
1i|T |
Ci, Ci :=
n
(x, y) 2 R|V|+1
| ⇢⇤
i (x)  y
o
.
numerical optimization we use a GPU-based implementation1
of a first-o
al-dual method [14]. The algorithm requires the orthogonal projectio
dual variables onto the sets C respectively K in every iteration. However
ection onto an epigraph of dimension |V| + 1 is di cult for large valu
We rewrite the constraints (v(x), w(x)) 2 Ci, 1  i  |T |, x 2 ⌦ as (n +
ensional epigraph constraints introducing variables ri
(x) 2 Rn
, si(x) 2
主問題の変数 双対問題の変数
• 𝑢に関する最適化は容易
• 𝑣, 𝑤, 𝑞の最適化（C, Kへの射影）が課題
𝑣
射影
射影とは・・・変数の動ける範囲が限定されている
場合に、範囲内に変数を移動させるステップ

34. 今回の問題を解くためには・・・
u:⌦!R
x2⌦
r to get rid of the pointwise maximum over ⇢⇤
i (v) in Eq. (8), we introdu
nal variables w(x) 2 R and additional constraints (v(x), w(x)) 2 C, x 2
w(x) attains the value of the pointwise maximum:
min
u:⌦!R|V|
max
(v,w):⌦!C
q:⌦!K
X
x2⌦
hu(x), v(x)i w(x) + hDiv q, ui, (1
the set C is given as
C =
1i|T |
Ci, Ci :=
n
(x, y) 2 R|V|+1
| ⇢⇤
i (x)  y
o
. (2
merical optimization we use a GPU-based implementation1
of a first-ord
dual method [14]. The algorithm requires the orthogonal projections
al variables onto the sets C respectively K in every iteration. However, th
ion onto an epigraph of dimension |V| + 1 is di cult for large values
rewrite the constraints (v(x), w(x)) 2 Ci, 1  i  |T |, x 2 ⌦ as (n + 1
u:⌦!R|V| q:⌦!K
x2⌦
der to get rid of the pointwise maximum over ⇢⇤
i (v) in Eq. (8), we introduce
tional variables w(x) 2 R and additional constraints (v(x), w(x)) 2 C, x 2 ⌦
at w(x) attains the value of the pointwise maximum:
min
u:⌦!R|V|
max
(v,w):⌦!C
q:⌦!K
X
x2⌦
hu(x), v(x)i w(x) + hDiv q, ui, (19)
e the set C is given as
C =
1i|T |
Ci, Ci :=
n
(x, y) 2 R|V|+1
| ⇢⇤
i (x)  y
o
. (20)
numerical optimization we use a GPU-based implementation1
of a first-order
al-dual method [14]. The algorithm requires the orthogonal projections of
dual variables onto the sets C respectively K in every iteration. However, the
ection onto an epigraph of dimension |V| + 1 is di cult for large values of
We rewrite the constraints (v(x), w(x)) 2 Ci, 1  i  |T |, x 2 ⌦ as (n + 1)-
nsional epigraph constraints introducing variables ri
(x) 2 Rn
, si(x) 2 R:
ri
(x)  si(x), ri
(x) = A>
i E>
i v(x), si(x) = w(x) hEibi, v(x)i. (21)
8 E. Laude, T. M¨ollenho↵, M. Moeller, J. Lellmann, D. Cremers
Proof. Follows from a calculation starting at the definition of the convex conju-
gate ⇤
. See Appendix A.
Interestingly, although in its original formulation (14) the set K has infinitely
many constraints, one can equivalently represent K by finitely many.
Proposition 3 The set K in equation (14) is the same as
K =
n
q 2 Rd⇥|V|
| Di
q S1  1, 1  i  |T |
o
, Di
q = QiD (TiD) 1
, (15)
where the matrices QiD 2 Rd⇥n
and TiD 2 Rn⇥n
are given as
QiD := qi1
qin+1
, . . . , qin
qin+1
, TiD := ti1
tin+1
, . . . , tin
tin+1
.
Proof. Similar to the analysis in [11], equation (14) basically states the Lipschitz
continuity of a piecewise linear function defined by the matrices q 2 Rd⇥|V|
.
Therefore, one can expect that the Lipschitz constraint is equivalent to a bound
• 𝑣、𝑤について最大化しつつ、領域Cに射影
• 𝑞について最大化しつつ、領域Kに射影

35. 射影について
al variables w(x) 2 R and additional constraints (v(x), w(x)) 2 C, x 2 ⌦
w(x) attains the value of the pointwise maximum:
min
u:⌦!R|V|
max
(v,w):⌦!C
q:⌦!K
X
x2⌦
hu(x), v(x)i w(x) + hDiv q, ui, (19)
he set C is given as
C =
1i|T |
Ci, Ci :=
n
(x, y) 2 R|V|+1
| ⇢⇤
i (x)  y
o
. (20)
erical optimization we use a GPU-based implementation1
of a first-order
dual method [14]. The algorithm requires the orthogonal projections of
variables onto the sets C respectively K in every iteration. However, the
on onto an epigraph of dimension |V| + 1 is di cult for large values of
rewrite the constraints (v(x), w(x)) 2 Ci, 1  i  |T |, x 2 ⌦ as (n + 1)-
onal epigraph constraints introducing variables ri
(x) 2 Rn
, si(x) 2 R:
x)  si(x), ri
(x) = A>
i E>
i v(x), si(x) = w(x) hEibi, v(x)i. (21)
quality constraints can be implemented using Lagrange multipliers. For
ection onto the set K we use an approach similar to [7, Figure 7].
Proof. Follows from a calculation starting at the definition of the convex conju-
gate ⇤
. See Appendix A.
Interestingly, although in its original formulation (14) the set K has infinitely
many constraints, one can equivalently represent K by finitely many.
Proposition 3 The set K in equation (14) is the same as
K =
n
q 2 Rd⇥|V|
| Di
q S1  1, 1  i  |T |
o
, Di
q = QiD (TiD) 1
, (15)
where the matrices QiD 2 Rd⇥n
and TiD 2 Rn⇥n
are given as
QiD := qi1
qin+1
, . . . , qin
qin+1
, TiD := ti1
tin+1
, . . . , tin
tin+1
.
Proof. Similar to the analysis in [11], equation (14) basically states the Lipschitz
continuity of a piecewise linear function defined by the matrices q 2 Rd⇥|V|
.
Therefore, one can expect that the Lipschitz constraint is equivalent to a bound
on the derivative. For the complete proof, see Appendix A.
Schatten-∞ Normへの射影
（Dの特異値の最大が１以下になるようなqを求める）
問題依存、パラボラ関数への射影など
𝑒𝑝𝑖 𝜌 + ∆R
∗

36. （参考）パラボラ関数への射影
• Convex Relaxation of Vectorial Problems with Coupled
Regularization (E. Strekalovskiy, A. Chambolle, D. Cremers), In
SIAM Journal on Imaging Sciences, volume 7, 2014.
CONVEX RELAXATION OF VECTORIAL PROBLEMS 333
B.2. Projection onto parabolas y ≥ α∥x∥2
2. Let α > 0. For x0 ∈ Rd and y0 ∈ R
consider the projection onto a parabola:
(B.4) arg min
x∈Rd, y∈R,
y≥α∥x∥2
2
(x − x0)2
2
+
(y − y0)2
2
.
If already y0 ≥ α∥x0∥2
2, the solution is (x, y) = (x0, y0). Otherwise, with a := 2α∥x0∥2,
b := 2
3(1 − 2αy0), and d := a2 + b3 set
(B.5) v :=
⎧
⎨
⎩
c − b
c with c =
3
a +
√
d if d ≥ 0,
2
√
−b cos 1
3 arccos a
√
−b
3 if d < 0.
If c = 0 in the first case, set v := 0. The solution is then given by
(B.6) x =
v
2α
x0
∥x0∥2
if x0 ̸= 0
0 else
, y = α∥x∥2
2.
Remark. In the case d < 0 it always holds that a
√
−b
3 ∈ [0, 1]. To ensure this also numeri-
cally, one should compute d by d = (a −
√
−b
3
)(a +
√
−b
3
) for b < 0.
Proof. First, for y0 ≥ α∥x0∥2
2 the projection is obviously (x, y) = (x0, y0). Otherwise, we

37. （参考） Schatten-∞ Normへの射影
• The Natural Total Variation Which Arises from Geometric Measure
Theory (B. Goldluecke, E. Strekalovskiy, D. Cremers), In SIAM Journal on
Imaging Sciences, volume 5, 2012.
in that color edges are preserved better. We also showed that TVJ can serve as a regularizer
in more general energy functionals, which makes it applicable to general inverse problems like
deblurring, zooming, inpainting, and superresolution.
7.1. Projection ΠS for TVS. Since each channel is treated separately, we can compute
the well-known projection for the scalar TV for each color channel. Let A ∈ Rn×m with rows
a1, . . . , an ∈ Rm. Then ΠS is defined rowwise as
(7.1) ΠS(ai) =
ai
max(1, |ai|2)
.
7.2. Projection ΠF for TVF . Let A ∈ Rn×m with elements aij ∈ R. From (2.8) we see
that we need to compute the projection onto the unit ball in Rn·m when (aij) is viewed as a
vector in Rn·m. Thus,
(7.2) ΠF (A) =
A
max 1, n
i=1
m
j=1 a2
ij
.
7.3. Projection ΠJ for TVJ . Let A ∈ Rn×m with singular value decomposition A =
UΣV T and Σ = diag(σ1, . . . , σm). We assume that the singular values are ordered with σ1
being the largest. If the sum of the singular values is less than or equal to one, A already lies
in co(En ⊗ Em). Otherwise, according to Theorem 3.18,
(7.3) Π(A) = UΣpV T
with Σp = diag(σp).
To compute the matrix V and the singular values, note that the Eigenvalue decomposition of
the m × m matrix AT A is given by V Σ2V T , which is more efficient to compute than the full
singular value decomposition since m < n. For images, m = 2, so there is even an explicit
formula available. We can now simplify the formula (7.3) to make the computation of U
unnecessary. Let Σ+ denote the pseudoinverse of Σ which is given by
(7.4) Σ+
= diag
1
σ1
, . . . ,
1
σk
, 0, . . . , 0 ,
where σk is the smallest nonzero singular value. Then U = AV Σ+, and from (7.3) we conclude
(7.5) Π(A) = AV Σ+
ΣpV T
.
For the special case of color images, where n = 3 and m = 2, the implementation of (7.5) is
detailed in Figure 7.
Appendix A. In this appendix we show explicitly how to compute the projection ΠK :

38. 実験：デノイジング
• 領域分け＋線形補完による最適化手法（右端）と、
提案手法（領域分け＋凸関数近似）の比較
• 領域数が少ないにも関わらず質の高い結果
Input image Unlifted Problem,
E = 992.50
Ours, |T | = 1,
|V| = 4,
E = 992.51
Ours, |T | = 6
|V| = 2 ⇥ 2 ⇥ 2
E = 993.52
Baseline,
|V| = 4 ⇥ 4 ⇥ 4,
E = 2255.81
Fig. 5: Convex ROF with vectorial TV. Direct optimization and proposed method
yield the same result. In contrast to the baseline method [11] the proposed ap-
proach has no discretization artefacts and yields a lower energy. The regulariza-
tion parameter is chosen as = 0.3.
Noisy input Ours, |T | = 1,
|V| = 4,
E = 2849.52
Ours, |T | = 6,
|V| = 2 ⇥ 2 ⇥ 2,
E = 2806.18
Ours, |T | = 48,
|V| = 3 ⇥ 3 ⇥ 3,
E = 2633.83
Baseline,
|V| = 4 ⇥ 4 ⇥ 4,
E = 3151.80
e purpose of this experiment is a proof of concept as our method i
overhead and convex problems can be solved via direct optimizatio
seen in Fig. 4 and Fig. 5, that the baseline method [11] has a str
as.
2 Denoising with Truncated Quadratic Dataterm
r images degraded with both, Gaussian and salt-and-pepper noise
e dataterm as ⇢(x, u(x)) = min 1
2 ku(x) I(x)k2
, ⌫ . We solve the

39. 実験：オプティカルフローΩImage 1 [8], |V| = 5 ⇥ 5,
0.67 GB, 4 min
aep = 2.78
[8], |V| = 11 ⇥ 11,
2.1 GB, 12 min
aep = 1.97
[8],
4.
Image 2 [11], |V| = 3 ⇥ 3,
0.67 GB, 0.35 min
aep = 5.44
[11], |V| = 5 ⇥ 5,
2.4 GB, 16 min
aep = 4.22
[11
5.
Ground truth Ours, |V| = 2 ⇥ 2,
0.63 GB, 17 min
aep = 1.28
Ours, |V| = 3 ⇥ 3,
1.9 GB, 34 min
aep = 1.07
Ou
4.
Fig. 7: We compute the optical flow using our met
Image 1 [8], |V| = 5 ⇥ 5,
0.67 GB, 4 min
aep = 2.78
[8], |V| = 11 ⇥ 11,
2.1 GB, 12 min
aep = 1.97
[8], |V| = 17 ⇥ 17,
4.1 GB, 25 min
aep = 1.63
[
Image 2 [11], |V| = 3 ⇥ 3,
0.67 GB, 0.35 min
aep = 5.44
[11], |V| = 5 ⇥ 5,
2.4 GB, 16 min
aep = 4.22
[11], |V| = 7 ⇥ 7,
5.2 GB, 33 min
aep = 2.65
Ground truth Ours, |V| = 2 ⇥ 2,
0.63 GB, 17 min
aep = 1.28
Ours, |V| = 3 ⇥ 3,
1.9 GB, 34 min
aep = 1.07
Ours, |V| = 4 ⇥ 4,
4.1 GB, 41 min
aep = 0.97
O
Fig. 7: We compute the optical flow using our method, the prod
Sublabel-Accurate Convex Relaxation of Vectorial Multil
Image 1 [8], |V| = 5 ⇥ 5,
0.67 GB, 4 min
aep = 2.78
[8], |V| = 11 ⇥ 11,
2.1 GB, 12 min
aep = 1.97
[8], |V| = 17 ⇥ 17,
4.1 GB, 25 min
aep = 1.63
[8], |
9.3
a
Image 2 [11], |V| = 3 ⇥ 3,
0.67 GB, 0.35 min
aep = 5.44
[11], |V| = 5 ⇥ 5,
2.4 GB, 16 min
aep = 4.22
[11], |V| = 7 ⇥ 7,
5.2 GB, 33 min
aep = 2.65
[11],
Out
Sublabel-Accurate Convex Relaxation of Vectorial Multilabel Energies
Image 1 [8], |V| = 5 ⇥ 5,
0.67 GB, 4 min
aep = 2.78
[8], |V| = 11 ⇥ 11,
2.1 GB, 12 min
aep = 1.97
[8], |V| = 17 ⇥ 17,
4.1 GB, 25 min
aep = 1.63
[8], |V| = 28 ⇥ 28,
9.3 GB, 60 min
aep = 1.39
Image 2 [11], |V| = 3 ⇥ 3,
0.67 GB, 0.35 min
aep = 5.44
[11], |V| = 5 ⇥ 5,
2.4 GB, 16 min
aep = 4.22
[11], |V| = 7 ⇥ 7,
5.2 GB, 33 min
aep = 2.65
[11], |V| = 9 ⇥ 9,
Out of memory.
ation of the energy instead of a
etween two input images I1, I2.
ding to the estimated maximum
dataterm is ⇢(x, v(x)) = kI2(x)
of the image gradient rI1(x).
d to the product space approach
ce dataterm using Lagrange mul-
memory as it has to store a con
linear one.
4.3 Optical Flow
We compute the optical flow v
The label space = [ d, d]2
is
displacement d 2 R between the
I1(x + v(x))k, and (x) is based
In Fig. 7 we compare the pr
[8]. Note that we implemented th
• 領域分け＋線形補完による最適化手法（右端）と、
提案手法（領域分け＋凸関数近似）の比較
• 領域数が少ないにも関わらず質の高い結果

40. まとめ
• 多次元のマルチラベル問題の一般解法を提案
• 凸関数で近似することで高精度な解を出力
• Vectorial Total Variationによる質の高い正則化