Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
画像処理論セミナー7-1-3
Search
Kuno Ayana
July 02, 2020
Education
41
0
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
画像処理論セミナー7-1-3
Kuno Ayana
July 02, 2020
More Decks by Kuno Ayana
See All by Kuno Ayana
アクセシビリティ、まだ完璧じゃないけど ── “今から”できること
kno3a87
2
1.4k
ぬるぬる動かせ! Riveでアニメーション実装🐾
kno3a87
1
2.1k
Dart 参戦!!静的型付き言語界の隠れた実力者
kno3a87
0
310
Flutterを言い訳にしない!アプリの使い心地改善テクニック5選🔥
kno3a87
3
850
iOS 18 がやってきた!
kno3a87
1
280
おうちハッカソン #2
kno3a87
0
170
ミクアカ成果報告会
kno3a87
0
82
SXSW2021
kno3a87
0
84
ミクアカ中間発表会
kno3a87
0
63
Other Decks in Education
See All in Education
SFプロトタイピング体験(滝中学校)
yusk1450
PRO
0
110
Portable & Reproducible Research Environments in the Age of AI Agents
denkiwakame
0
610
Soluciones al examen de Geografía 2026. JULIO (Convocatoria Extraordinaria)
juanmartin2026
1
17k
私たちはなんでテストするんだっけ? Ver.東北IT物産展2026 in 会津若松
camel_404
3
140
Human-AI Interaction - Lecture 11 - Next Generation User Interfaces (4018166FNR)
signer
PRO
0
1.1k
1人 × AI、1か月でここまで作れる ー 数年前の外注換算3.8〜7.4億円・241〜379人月分の作業を、AI費用 約10万円・31日で
frievea
0
330
BITCOIN : Les fondamentaux !
rlifchitz
0
220
Lectura 2 (PIT : Python Basico)
robintux
0
400
教育現場から見た Ruby on Rails
yasslab
PRO
0
240
Visionary Initiative: Materials-Positive Society 「モノの進化をポジティブな社会の原動力に」|Science Tokyo(東京科学大学)
sciencetokyo
PRO
0
920
2026年度春学期 統計学 第9回 確からしさを記述する ー 確率 (2026. 5. 28)
akiraasano
PRO
0
170
Geografía y Fútbol: Chattanooga Geografía del Búnker de La Roja.
juanmartin2026
1
12k
Featured
See All Featured
Leo the Paperboy
mayatellez
8
2.1k
End of SEO as We Know It (SMX Advanced Version)
ipullrank
3
4.3k
Rebuilding a faster, lazier Slack
samanthasiow
85
9.6k
Building the Perfect Custom Keyboard
takai
2
830
More Than Pixels: Becoming A User Experience Designer
marktimemedia
3
480
Fantastic passwords and where to find them - at NoRuKo
philnash
52
3.8k
Technical Leadership for Architectural Decision Making
baasie
3
470
Evolving SEO for Evolving Search Engines
ryanjones
0
250
Primal Persuasion: How to Engage the Brain for Learning That Lasts
tmiket
0
400
RailsConf 2023
tenderlove
30
1.5k
Taking LLMs out of the black box: A practical guide to human-in-the-loop distillation
inesmontani
PRO
3
2.3k
The AI Search Optimization Roadmap by Aleyda Solis
aleyda
1
6k
Transcript
,VOP"ZBOB σΟδλϧը૾ॲཧ ٯϑΟϧλɾΟʔφϑΟϧλʹΑΔը૾෮ݩ
લճͷ෮़ɿ΅͚ɾͿΕͱ ࣍ݩσϧλؔ δ(x, y) ྼԽը૾ g(x, y) ݪը૾ f(x, y)
લճͷ෮़ɿ֦͕ΓؔͷϞσϧԽ ΅͚ͷ֦͕ΓؔˠΨεͱۙࣅ ͿΕͷ֦͕ΓؔˠͿΕͷํВʹͷΈ෯XʹҰ࣍ݩͰ͕͍ͬͯΔؔͱۙࣅ
ը૾Λ෮ݩ͢ΔۭؒϑΟϧλΛߟ͑Δ g(x, y) = f(x, y) * h(x, y) ྼԽը૾
ݪը૾ ֦͕Γؔ ֦͕Γ͕ؔطͷ߹ ٯϑΟϧλΛྼԽը૾ʹదԠ͢Δ͜ͱͰݪը૾͕ٻ·Δ
ը૾Λ෮ݩ͢ΔۭؒϑΟϧλΛߟ͑Δ g(x, y) = f(x, y) * h(x, y) ྼԽը૾
ݪը૾ G(u, v) = F(u, v)H(u, v) ϑʔϦΤม 'ྼԽը૾ 'ݪը૾ ϑΟϧλ ֦͕Γؔ ֦͕Γ͕ؔطͷ߹ ٯϑΟϧλΛྼԽը૾ʹదԠ͢Δ͜ͱͰݪը૾͕ٻ·Δ
ը૾Λ෮ݩ͢ΔۭؒϑΟϧλΛߟ͑Δ G(u, v) = F(u, v)H(u, v) 'ྼԽը૾ 'ݪը૾ ϑΟϧλ
֦͕Γ͕ؔطͷ߹ ٯϑΟϧλΛྼԽը૾ʹదԠ͢Δ͜ͱͰݪը૾͕ٻ·Δ
ը૾Λ෮ݩ͢ΔۭؒϑΟϧλΛߟ͑Δ G(u, v) = F(u, v)H(u, v) 'ྼԽը૾ 'ݪը૾ ϑΟϧλ
֦͕Γ͕ؔطͷ߹ ٯϑΟϧλΛྼԽը૾ʹదԠ͢Δ͜ͱͰݪը૾͕ٻ·Δ K(u, v) K(u, v)
ը૾Λ෮ݩ͢ΔۭؒϑΟϧλΛߟ͑Δ G(u, v) = F(u, v)H(u, v) 'ྼԽը૾ 'ݪը૾ ϑΟϧλ
1 H(u, v) ٯϑΟϧλ ֦͕Γ͕ؔطͷ߹ ٯϑΟϧλΛྼԽը૾ʹదԠ͢Δ͜ͱͰݪը૾͕ٻ·Δ 1 H(u, v)
ը૾Λ෮ݩ͢ΔۭؒϑΟϧλΛߟ͑Δ G(u, v) = F(u, v)H(u, v) 'ྼԽը૾ 'ݪը૾ ϑΟϧλ
1 H(u, v) ٯϑΟϧλ ֦͕Γ͕ؔطͷ߹ ٯϑΟϧλΛྼԽը૾ʹదԠ͢Δ͜ͱͰݪը૾͕ٻ·Δ 1 H(u, v)
ը૾Λ෮ݩ͢ΔۭؒϑΟϧλΛߟ͑Δ G(u, v) = F(u, v) 'ྼԽը૾ 'ݪը૾ ֦͕Γ͕ؔطͷ߹ ٯϑΟϧλΛྼԽը૾ʹదԠ͢Δ͜ͱͰݪը૾͕ٻ·Δ
1 H(u, v)
ը૾Λ෮ݩ͢ΔۭؒϑΟϧλΛߟ͑Δ G(u, v) = F(u, v) 'ྼԽը૾ 'ݪը૾ ֦͕Γ͕ؔطͷ߹ ٯϑΟϧλΛྼԽը૾ʹదԠ͢Δ͜ͱͰݪը૾͕ٻ·Δ
1 H(u, v)
ը૾Λ෮ݩ͢ΔۭؒϑΟϧλΛߟ͑Δ G(u, v) = F(u, v) 'ྼԽը૾ 'ݪը૾ ϑʔϦΤٯม g(x,
y) = f(x, y) ྼԽը૾ ݪը૾ ֦͕Γ͕ؔطͷ߹ ٯϑΟϧλΛྼԽը૾ʹదԠ͢Δ͜ͱͰݪը૾͕ٻ·Δ 1 H(u, v)
ը૾Λ෮ݩ͢ΔۭؒϑΟϧλΛߟ͑Δ G(u, v) = F(u, v)H(u, v) 'ྼԽը૾ 'ݪը૾ ϑΟϧλ
1 H(u, v) ٯϑΟϧλ ֦͕Γ͕ؔطͷ߹ ٯϑΟϧλΛྼԽը૾ʹదԠ͢Δ͜ͱͰݪը૾͕ٻ·Δ ͘͠ݶΓͳ͘ʹ͍ۙͩͬͨΒʁ
ը૾Λ෮ݩ͢ΔۭؒϑΟϧλΛߟ͑Δ G(u, v) = F(u, v)H(u, v) 'ྼԽը૾ 'ݪը૾ ϑΟϧλ
1 H(u, v) ٯϑΟϧλ ֦͕Γ͕ؔطͷ߹ ٯϑΟϧλΛྼԽը૾ʹదԠ͢Δ͜ͱͰݪը૾͕ٻ·Δ ͘͠ݶΓͳ͘ʹ͍ۙͩͬͨΒʁ ൃࢄͯ͠͠·͏ʂ
ը૾Λ෮ݩ͢ΔۭؒϑΟϧλΛߟ͑Δ G(u, v) = F(u, v)H(u, v) + N(u, v)
ൃࢄ͢ΔͱϊΠζ͕૿෯ͯ͠͠·͏ ˠ) V W ͕ʹ͍ۙͱ͖ʹൃࢄ͠ͳ͍ϑΟϧλΛߟ͑Δඞཁ͕͋Δ 'ϊΠζ
ը૾Λ෮ݩ͢ΔۭؒϑΟϧλΛߟ͑Δ Kw (u, v) = 1 H(u, v) ෮ݩը૾ɹɹɹͱݪը૾ɹɹɹͷޡࠩΛ࠷খʹ͢ΔΑ͏ͳϑΟϧλ Kw
(u, v) ̂ f(x, y) f(x, y)
ը૾Λ෮ݩ͢ΔۭؒϑΟϧλΛߟ͑Δ Kw (u, v) = 1 H(u, v) |H(u, v)|2
|H(u, v)|2 + |N(u, v)|2 /|F(u, v)|2 ෮ݩը૾ɹɹɹͱݪը૾ɹɹɹͷޡࠩΛ࠷খʹ͢ΔΑ͏ͳϑΟϧλ Kw (u, v) ̂ f(x, y) f(x, y)
ը૾Λ෮ݩ͢ΔۭؒϑΟϧλΛߟ͑Δ Kw (u, v) = 1 H(u, v) |H(u, v)|2
|H(u, v)|2 + |N(u, v)|2 /|F(u, v)|2 ෮ݩը૾ɹɹɹͱݪը૾ɹɹɹͷޡࠩΛ࠷খʹ͢ΔΑ͏ͳϑΟϧλ Kw (u, v) ̂ f(x, y) f(x, y) ΟʔφϑΟϧλ
ը૾Λ෮ݩ͢ΔۭؒϑΟϧλΛߟ͑Δ Kw (u, v) = 1 H(u, v) |H(u, v)|2
|H(u, v)|2 + |N(u, v)|2 /|F(u, v)|2 ෮ݩը૾ɹɹɹͱݪը૾ɹɹɹͷޡࠩΛ࠷খʹ͢ΔΑ͏ͳϑΟϧλ Kw (u, v) ̂ f(x, y) f(x, y) ϊΠζ͕ͷ߹
ը૾Λ෮ݩ͢ΔۭؒϑΟϧλΛߟ͑Δ Kw (u, v) = 1 H(u, v) |H(u, v)|2
|H(u, v)|2 + |N(u, v)|2 /|F(u, v)|2 ෮ݩը૾ɹɹɹͱݪը૾ɹɹɹͷޡࠩΛ࠷খʹ͢ΔΑ͏ͳϑΟϧλ Kw (u, v) ̂ f(x, y) f(x, y) ϊΠζ͕ͷ߹ ͕͜͜ʹͳΔͷͰ ٯϑΟϧλͱಉ༷ʹΔ
ը૾Λ෮ݩ͢ΔۭؒϑΟϧλΛߟ͑Δ Kw (u, v) = 1 H(u, v) |H(u, v)|2
|H(u, v)|2 + |N(u, v)|2 /|F(u, v)|2 ෮ݩը૾ɹɹɹͱݪը૾ɹɹɹͷޡࠩΛ࠷খʹ͢ΔΑ͏ͳϑΟϧλ Kw (u, v) ̂ f(x, y) f(x, y) ͍͍ͩͨϊΠζݪը૾ະ దͳఆϵΛஔ͘͜ͱ͕ଟ͍
ը૾Λ෮ݩ͢ΔۭؒϑΟϧλΛߟ͑Δ Kw (u, v) = 1 H(u, v) |H(u, v)|2
|H(u, v)|2 + Γ ෮ݩը૾ɹɹɹͱݪը૾ɹɹɹͷޡࠩΛ࠷খʹ͢ΔΑ͏ͳϑΟϧλ Kw (u, v) ̂ f(x, y) f(x, y)
ϵΛมԽͤ͞Δ Kw (u, v) = 1 H(u, v) |H(u, v)|2
|H(u, v)|2 + Γ ਤ
ϵΛมԽͤ͞Δ Kw (u, v) = 1 H(u, v) |H(u, v)|2
|H(u, v)|2 + Γ ਤ ൃࢄͯ͠͠·͍ըૉ͕ൃࢄ͍ͯ͠Δ θϩΛؚΜͰ͍ΔͨΊ
ϵΛมԽͤ͞Δ Kw (u, v) = 1 H(u, v) |H(u, v)|2
|H(u, v)|2 + Γ ਤ ൃࢄͯ͠͠·͍ըૉ͕ൃࢄ͍ͯ͠Δ θϩΛؚΜͰ͍ΔͨΊ ൃࢄ͍ͯ͠ͳ͍
ϵΛมԽͤ͞Δ Kw (u, v) = 1 H(u, v) |H(u, v)|2
|H(u, v)|2 + Γ ਤ
ϵΛมԽͤ͞Δ Kw (u, v) = 1 H(u, v) |H(u, v)|2
|H(u, v)|2 + Γ ਤ Ӷʹ෮ݩ͞ΕΔ ϊΠζ૿෯͢Δ ϊΠζ૿෯͞Εͳ͍ ΅͚ɾͿΕͷ෮ݩ͕͍
ϵΛมԽͤ͞Δ Kw (u, v) = 1 H(u, v) |H(u, v)|2
|H(u, v)|2 + Γ ਤ Ӷʹ෮ݩ͞ΕΔ ϊΠζ૿෯͢Δ ϊΠζ૿෯͞Εͳ͍ ΅͚ɾͿΕͷ෮ݩ͕͍ ϵ͕େ͖͘ͳΔͱ͕େ͖͘ͳΔͷͰ
ϵΛมԽͤ͞Δ Kw (u, v) = 1 H(u, v) |H(u, v)|2
|H(u, v)|2 + Γ ਤ Ӷʹ෮ݩ͞ΕΔ ϊΠζ૿෯͢Δ ϊΠζ૿෯͞Εͳ͍ ΅͚ɾͿΕͷ෮ݩ͕͍ ϵ͕େ͖͘ͳΔͱ͕େ͖͘ͳΔͷͰ ͜͜ͷ͕খ͘͞ͳͬͯ͋·ΓϑΟϧλ͕ޮ͔ͳ͘ͳΔ