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7-1,2
Search
Kuno Ayana
June 19, 2020
Education
0
31
画像処理論セミナー7-1,2
Kuno Ayana
June 19, 2020
Tweet
Share
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
1k
ぬるぬる動かせ! Riveでアニメーション実装🐾
kno3a87
1
1.8k
Dart 参戦!!静的型付き言語界の隠れた実力者
kno3a87
0
250
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
820
iOS 18 がやってきた!
kno3a87
1
250
おうちハッカソン #2
kno3a87
0
150
ミクアカ成果報告会
kno3a87
0
62
SXSW2021
kno3a87
0
69
ミクアカ中間発表会
kno3a87
0
46
Other Decks in Education
See All in Education
2025年の本当に大事なAI動向まとめ
frievea
1
190
Information Architectures - Lecture 2 - Next Generation User Interfaces (4018166FNR)
signer
PRO
1
1.8k
Gitの中身 / 03-a-git-internals
kaityo256
PRO
0
110
AIで日本はどう進化する? 〜キミが生きる2035年の地図〜
behomazn
0
130
Web Search and SEO - Lecture 10 - Web Technologies (1019888BNR)
signer
PRO
2
3.1k
0121
cbtlibrary
0
140
HCI Research Methods - Lecture 7 - Human-Computer Interaction (1023841ANR)
signer
PRO
0
1.4k
Measuring your measuring
jonoalderson
2
730
Gluon Recruit Deck
gluon
0
150
Adobe Express
matleenalaakso
2
8.2k
環境・社会理工学院(建築学系)大学院説明会 2026|東京科学大学(Science Tokyo)
sciencetokyo
PRO
0
400
AIでキミの未来はどう変わる?
behomazn
0
120
Featured
See All Featured
Taking LLMs out of the black box: A practical guide to human-in-the-loop distillation
inesmontani
PRO
3
2.1k
Prompt Engineering for Job Search
mfonobong
0
180
Being A Developer After 40
akosma
91
590k
16th Malabo Montpellier Forum Presentation
akademiya2063
PRO
0
65
Tell your own story through comics
letsgokoyo
1
830
Mobile First: as difficult as doing things right
swwweet
225
10k
Git: the NoSQL Database
bkeepers
PRO
432
66k
Sam Torres - BigQuery for SEOs
techseoconnect
PRO
0
210
What's in a price? How to price your products and services
michaelherold
247
13k
Responsive Adventures: Dirty Tricks From The Dark Corners of Front-End
smashingmag
254
22k
Put a Button on it: Removing Barriers to Going Fast.
kastner
60
4.2k
Redefining SEO in the New Era of Traffic Generation
szymonslowik
1
240
Transcript
,VOP"ZBOB σΟδλϧը૾ॲཧ ΅͚ɾͿΕը૾ͷ෮ݩ
ը૾ͷྼԽ g(x, y) = f(x, y) * h(x, y) ྼԽը૾
ݪը૾ ϑΟϧλؔ Χϝϥͷয΅͚ͿΕʹΑΔྼԽ ˠݪը૾ʹۭؒϑΟϧλϦϯάॲཧΛࢪͨ͠ը૾ͱߟ͑ΒΕΔ
ը૾ͷྼԽ g(x, y) = f(x, y) * h(x, y) ྼԽը૾
ݪը૾ ϑΟϧλؔ Χϝϥͷয΅͚ͿΕʹΑΔྼԽ ˠݪը૾ʹۭؒϑΟϧλϦϯάॲཧΛࢪͨ͠ը૾ͱߟ͑ΒΕΔ
ը૾ͷྼԽ ࣍ݩσϧλؔ δ(x, y) ྼԽը૾ g(x, y) ݪը૾ f(x, y)
g(x, y) = f(x, y) * h(x, y) + n(x,
y) ྼԽը૾ ݪը૾ ֦͕Γؔ ը૾Λ෮ݩ͍ͨ͠ʂ ˠϊΠζʹΑΔྼԽߟྀ͠ͳ͍͚ͯ͘ͳ͍ ϊΠζ ը૾ͷ෮ݩ
g(x, y) = f(x, y) * h(x, y) + n(x,
y) ྼԽը૾ ݪը૾ ֦͕Γؔ ϊΠζ ը૾ͷ෮ݩ ը૾Λ෮ݩ͍ͨ͠ʂ ˠ·֦͕ͣΓؔΛΓ͍ͨ
֦͕ΓؔΛΓ͍ͨ h(x, y) * δ(x, y) = h(x, y) ֦͕Γؔ
࣍ݩσϧλؔ ݪը૾ˎ ࣍ݩσϧλؔ ݪը૾ ˠ֦͕Γؔˎ ࣍ݩσϧλؔ ֦͕Γؔ ֦͕Γؔ
֦͕ΓؔΛΓ͍ͨ h(x, y) * δ(x, y) = h(x, y) ֦͕Γؔ
࣍ݩσϧλؔ ݪը૾ˎ ࣍ݩσϧλؔ ݪը૾ ˠ֦͕Γؔˎ ࣍ݩσϧλؔ ֦͕Γؔ ֦͕Γؔ ݪը૾͕શͳʮʯͰ͋Ε ྼԽը૾ͱͯ͠ϑΟϧλ͕ؔಘΒΕΔʂ
֦͕ΓؔͷϞσϧԽ ΅͚ͷ֦͕Γؔ ˠΨεͱۙࣅ͍ͯ͠Δ h(x, y) = 1 2πσ2 exp(− x2
+ y2 2σ2 )
֦͕ΓؔͷϞσϧԽ ͿΕͷ֦͕Γؔ ˠͿΕͷํВʹͷΈ෯XʹҰ࣍ݩͰ͕͍ͬͯΔؔͱۙࣅ͍ͯ͠Δ |xcosθ + ycosθ| ≦ w 2 ∩
xcosθ − ycosθ = 0 hθ = { 1 w 0 ͦΕҎ֎
֦͕Γؔͷύϥϝʔλ ֦͕ΓؔΛදͨ͢Ίʹ ɾ΅͚ͷఔМ ɾͿΕͷํВɼͿΕͷఔXɹͷύϥϝʔλΛఆΊΔඞཁ͕͋Δ ɾϨϯζͷಛੑ ɾγϟολʔεϐʔυɼҠಈɾํ ͕Θ͔͍ͬͯΕύϥϝʔλΛ͋Β͔͡ΊܾΊΒΕΔ
֦͕Γؔͷύϥϝʔλ ֦͕ΓؔΛදͨ͢Ίʹ ɾ΅͚ͷఔМ ɾͿΕͷํВɼͿΕͷఔXɹͷύϥϝʔλΛఆΊΔඞཁ͕͋Δ ɾϨϯζͷಛੑ ɾγϟολʔεϐʔυɼҠಈɾํ ͕Θ͔͍ͬͯΕύϥϝʔλΛ͋Β͔͡ΊܾΊΒΕΔ ͍͍ͩͨະ
֦͕Γؔͷύϥϝʔλਪͷྫ ֦͕Γؔͷ࣍ݩϑʔϦΤมɹɹɹपྖҬͰपظతʹʹͳΔ ˠྼԽը૾ͷ࣍ݩϑʔϦΤมɹɹɹಉ༷ʹपྖҬͰपظతʹʹͳΔ H(u, v) G(u, v) G(u, v)
֦͕Γؔͷύϥϝʔλਪͷྫ ֦͕Γؔͷ࣍ݩϑʔϦΤมɹɹɹपྖҬͰपظతʹʹͳΔ ˠྼԽը૾ͷ࣍ݩϑʔϦΤมɹɹɹಉ༷ʹपྖҬͰपظతʹʹͳΔ H(u, v) G(u, v) ྼԽը૾ͷ࣍ݩϑʔϦΤมɹɹɹ͕ʹͳΔपظΛݕग़͢Δ͜ͱͰ ϞσϧύϥϝʔλΛਪఆͰ͖Δʂ G(u,
v)