Slide 1

Slide 1 text

4.2 અ Τοδ ઒ౡوେ June 11, 2018 ిؾ௨৴େֶ ঙ໺ݚڀࣨ B4

Slide 2

Slide 2 text

໨࣍ 1. Τοδͷݕग़ 2. Τοδͷ࿈݁ 2

Slide 3

Slide 3 text

Τοδͷݕग़

Slide 4

Slide 4 text

Τοδͷݕग़ ྠֲઢͳͲͷΤοδ͸͖ΘΊͯଟ͘ͷ৘ใΛؚΉ ਓखʹΑΔΤοδݕग़ (ਤ 4.31) ˠ͜ΕΛύιίϯ༷ʹ΍Β͍ͤͨ 3

Slide 5

Slide 5 text

Τοδͷݕग़ ୯७ͳΤοδͷݕग़ํ๏ɿΤοδΛٸܹͳً౓มԽͱͯ͠ѻ͏ ˠً౓஋ͷޯ഑Λߟ͑Δ I(x) ΛϐΫηϧ x = (x, y)⊤ ্ͷً౓஋ͱ͢Δͱɼً౓ޯ഑ J(x) ͸ J(x) = ∇I(x) = ( ∂I ∂x , ∂I ∂y ) (x) (4.19) 4

Slide 6

Slide 6 text

Τοδͷݕग़ ϕΫτϧ J(x) ͷ • ޲͖ɿً౓ؔ਺ͷ࠷ٸޯ഑ํ޲ • େ͖͞ɿً౓ؔ਺ͷมԽ౓߹͍ 5

Slide 7

Slide 7 text

Τοδͷݕग़ ߴप೾੒෼ʹ͸ϊΠζ͕ଟ͍ ˠϩʔύεϑΟϧλͰฏ׈Խ͔ͯ͠Βޯ഑Λܭࢉ ローパス フィルタ 6

Slide 8

Slide 8 text

Τοδͷݕग़ ϑΟϧλద༻ޙ΋ޯ഑ͷ޲͖͕ਖ਼͘͠อଘ͞Ε͍ͯͯ΄͍͠ ˠԁܗͷϑΟϧλ ෼཭ՄೳͳԁܗϑΟϧλ͸Ψ΢εϑΟϧλͷΈ (3.2 અɼਤ 3.14) ˠΤοδݕग़ͷͨΊͷϩʔύεϑΟϧλ͸Ψ΢γΞϯ͕ఆ൪ 7

Slide 9

Slide 9 text

Τοδͷݕग़ ඍ෼͸ઢܗԋࢉͰ͋ΔͷͰଞͷϑΟϧλԋࢉͱՄ׵ Ψ΢εϑΟϧλؔ਺Λ Gσ(x) = 1 2πσ2 exp ( − x2 + y2 2σ2 ) ͱ͢Δ ฏ׈Խޙͷը૾ͷޯ഑Λ Jσ(x) ͱॻ͘ͱɼ Jσ(x) = ∇[Gσ(x) ∗ I(x)] = [∇Gσ(x)] ∗ I(x) (4.20) ͱͳΓɼΨ΢εϑΟϧλؔ਺ͷඍ෼ͱͷͨͨΈࠐΈͰදݱͰ͖Δ 8

Slide 10

Slide 10 text

Τοδͷݕग़ Ψ΢εϑΟϧλؔ਺ͷඍ෼ͷධՁ ∇Gσ(x) = ( ∂ ∂x , ∂ ∂y )⊤ Gσ(x) = ( ∂ ∂x , ∂ ∂y )⊤ 1 2πσ2 exp ( − x2 + y2 2σ2 ) = 1 σ2 (−x, − y)⊤ 1 2πσ2 exp ( − x2 + y2 2σ2 ) ((4.21) ࣜͱ߹Θͳ͍͕ͨͿΜ͜ΕͰ͍͋ͬͯΔ) 9

Slide 11

Slide 11 text

Τοδͷݕग़ thinning ΤοδΛ 1 ըૉͷଠ͞Ͱදݱ͍ͨ͠৔߹͕ଟ͍ (ࡉઢԽ; thinning) (ը૾͸ [1] ΑΓ) 10

Slide 12

Slide 12 text

Τοδͷݕग़ thinning ʮΤοδʹରͯ͠ਨ௚ͳํ޲ͷޯ഑ڧ౓͕࠷େʹͳΔ࠲ඪʯΛٻ ΊΕ͹Α͍ ˠً౓ͷ 2 ֊ඍ෼ (ϥϓϥγΞϯ) Λߟ͑Ε͹Αͦ͞͏ͩ ͜ͷ 2 ֊ඍ෼ͷ஋ Sσ(x) ͸ɼ∇2 = ∇ · ∇(= div grad) ΑΓ Sσ(x) = ∇ · Jσ(x) = [∇2Gσ(x)] ∗ I(x) (4.22) 11

Slide 13

Slide 13 text

Τοδͷݕग़ thinning Ψ΢εϑΟϧλͷϥϓϥγΞϯͷධՁ ∇2Gσ(x) = ∇ · [ 1 σ2 (−x, − y)⊤ 1 2πσ2 exp ( − x2 + y2 2σ2 )] = ∂ ∂x [ − x 2πσ4 exp ( − x2 + y2 2σ2 )] + ∂ ∂y [ − y 2πσ4 exp ( − x2 + y2 2σ2 )] = 1 2πσ2 ( x2 + y2 − 2σ2 σ4 ) exp ( − x2 + y2 2σ2 ) 12

Slide 14

Slide 14 text

Τοδͷݕग़ thinning ∇2Gσ(x) ͷ܎਺Λແࢹˠ LoG(Laplacian of Gaussian) ϑΟϧλ LoG(x) = ( x2 + y2 − 2σ2 σ4 ) exp ( − x2 + y2 2σ2 ) 13

Slide 15

Slide 15 text

Τοδͷݕग़ thinning Sσ(x) ͷූ߸͕มԽ ˠ૬ରతͳ໌Δ͕͞มԽ Sσ(x) ͷθϩަࠩ఺Λ୳ͤ͹ Α͍ 14

Slide 16

Slide 16 text

Τοδͷݕग़ thinning sign(Sσ(xi)) ̸= sign(Sσ(xj)) ͱͳΔྡ઀ϐΫηϧ xi, xj ͓Αͼθ ϩަࠩ఺ xz Λ୳͢ Sσ(xi) ͱ Sσ(xj) ͱΛ݁Ϳઢ෼͕θϩͱަࠩ͢Δ఺ xz ΛٻΊΔ 15

Slide 17

Slide 17 text

Τοδͷݕग़ thinning Sσ(xj) − Sσ(xi) xj − xi (xz − xi) + Sσ(xi) = 0 ∴ xz = xiSσ(xj) + xjSσ(xi) Sσ(xj) + Sσ(xi) ͕ಘΒΕΔɽ3 ࣍ݩҎ্ͷ৔߹΋ಉ༷ʹ xz = xiSσ(xj) + xjSσ(xi) Sσ(xj) + Sσ(xi) (4.25) Ͱ͋Δ 16

Slide 18

Slide 18 text

Τοδͷݕग़ εέʔϧબ୒ͱϘέྔਪఆ LoG ʹద౰ͳ σ ΛઃఆˠӶ͍/ಷ͍ΤοδΛநग़ (ਤ 4.32, (b), (c)) 17

Slide 19

Slide 19 text

Τοδͷݕग़ εέʔϧબ୒ͱϘέྔਪఆ ޿͍ײ౓ͰΤοδΛͱΓ͍ͨͳΒʁ ˠεέʔϧεϖʔεͷΞϓϩʔν 1. ͍͔ͭ͘ͷ σ Λ༻ҙ 2. ͦΕͧΕͷ σ ʹ͍ͭͯޯ഑ ͱ 2 ֊ඍ෼Λܭࢉ 3. ҆ఆʹΤοδΛݕग़Ͱ͖Δ ࠷খͷ σ Λબ୒ɼͦΕΑΓ େ͖͍ σ Ͱݕग़͞ΕͨΤο δΛ௥Ճ 18

Slide 20

Slide 20 text

Τοδͷݕग़ εέʔϧબ୒ͱϘέྔਪఆ ޿͍ σ ͰΤοδΛநग़ (ਤ 4.32, (f)) 19

Slide 21

Slide 21 text

Τοδͷݕग़ Χϥʔը૾ͰͷΤοδݕग़ Χϥʔը૾Ͱ΋Τοδݕग़Λ͍ͨ͠ ୯७ʹً౓ޯ഑ΛݟΔͱɼ౳ً౓৭ؒͷΤοδΛݕग़Ͱ͖ͳ͍ ղܾҊ 1ɿRGB ֤੒෼͝ͱʹً౓ޯ഑Λܭࢉ͢Δ • ֤৭Ͱූ߸ͷҟͳΔޯ഑͕ग़Δͱɼ୯७ͳ଍͠߹ΘͤͰ͸૬ ࡴ͕ى͜Δ ղܾҊ 2ɿ֤ըૉͷपลͰہॴతͳ౷ܭྔΛ͍Ζ͍Ζௐ΂Δ • ୯७ͳً౓ɾ໌౓ɾ৭͚ͩͰͳ͘ɼςΫενϟͷมԽͳͲ΋ ଊ͑ΒΕΔ 20

Slide 22

Slide 22 text

Τοδͷݕग़ ਤ 4.33ɽBGɿ໌౓ɼCGɿ৭ɼTGɿςΫενϟ 21

Slide 23

Slide 23 text

Τοδͷ࿈݁

Slide 24

Slide 24 text

Τοδͷ࿈݁ நग़͞ΕͨΤοδΛ࿈݁ͯ͠Ұܨ͗ʹ͍ͨ͠ thinning ͞ΕͨΤοδͷըૉ৘ใΛ͍࣋ͬͯΔͱָ ˠ͍ۙ୺఺Λ୳ࡧͯ͠ܨ͛͹Α͍ ΤοδΛ࿈݁͢ΔͱΑΓѹॖͨ͠දݱ͕ՄೳʹͳΔ 22

Slide 25

Slide 25 text

Τοδͷ࿈݁ νΣΠϯίʔυ 8 ͭͷํ֯ (N, NE, E, SE, S, SW, W, NW) Λ 3bit ͰίʔυԽ (ਤ 4.34) 23

Slide 26

Slide 26 text

Τοδͷ࿈݁ νΣΠϯίʔυ νΣΠϯίʔυͰͷΤϯίʔυޙɼϥϯϨϯάεූ߸Ͱ͞Βʹѹ ॖͰ͖Δ ϥϯϨϯάεූ߸ ܁Γฦ͠ͷจࣈΛͦͷճ਺Ͱදݱ AAAABBBCCCCC ˠ A4B3C5 24

Slide 27

Slide 27 text

Τοδͷ࿈݁ arc-length parameterization ʮހʯͷ௕͞ͱΤοδ࠲ඪΛ༻͍ͯදݱ (ਤ 4.35) 1. x0 = (1, 0.5)⊤ ͔Βελʔτ 2. s = 0 ʹ x0 ͷ࠲ඪΛͦΕͧΕϓϩοτ 3. x1 = (2, 0.5)⊤ 4. s = ∥x1 − x0∥ = 1 ʹ x1 ͷ࠲ඪΛͦΕͧΕϓϩοτ 5. ࢝఺ʹ໭Δ·Ͱ܁Γฦ͢ 25

Slide 28

Slide 28 text

Τοδͷ࿈݁ arc-length parameterization Q. Կ͕͏Ε͍͠ͷ͔ʁ A. Ϛονϯά΍ฏ׈ԽͳͲͷॲཧ͕༰қʹͳΔ ܗঢ়ͷࣅͨΤοδΛߟ͑Δ (ਤ 4.36) 26

Slide 29

Slide 29 text

Τοδͷ࿈݁ arc-length parameterization 1. Τοδͷ࠲ඪͷฏۉ஋ ¯ x0 = ∫ S x(s)ds Λݮࢉ 2. s Λ 0 ∼ S ͔Β 0 ∼ 1 ʹਖ਼نԽ 3. ͦΕͧΕʹ͍ͭͯϑʔϦΤม׵ 27

Slide 30

Slide 30 text

Τοδͷ࿈݁ arc-length parameterization ΋ͱͷΤοδಉ͕࢜εέʔϦϯάͱճసͷҧ͍͔͠ͳ͍ ˠϑʔϦΤม׵ͷ݁Ռ͸ڧ౓ͱҐ૬ͷζϨ͔͠ҟͳΒͳ͍͸ͣ (։࢝఺͕ҟͳΔͱઢܗͷҐ૬ͷζϨ΋ग़Δ) 28

Slide 31

Slide 31 text

Τοδͷ࿈݁ arc-length parameterization ཭ࢄԽ࣌ʹੜ͡ΔϊΠζͷฏ׈Խʹ΋༗ޮ ͔͠͠ී௨ʹฏ׈ԽϑΟϧλΛ͔͚Δͱॖখͯ͠ฏ׈Խ͞ΕΔ ਤ 4.37(a), ԁͷ൒ܘ͕ॖখ͍ͯ͠Δ 29

Slide 32

Slide 32 text

Τοδͷ࿈݁ arc-length parameterization 2 ֊ඍ෼ʹجͮ͘Φϑηοτ߲Λ଍͔͢ɼΑΓେ͖ͳ (ͦ͢ͷ޿ ͍ʁ) ฏ׈ԽϑΟϧλΛ༻͍Δ ਤ 4.37(b) 30

Slide 33

Slide 33 text

·ͱΊ • άϨʔεέʔϧը૾Ͱ͸ً౓ޯ഑ͰΤοδΛݕग़ ϊΠζআڈ΋ಉ࣌ʹߦ͏ͨΊʹΨ΢γΞϯϑΟϧλͷ 1 ֊ඍ ෼ͱͨͨΈࠐΉ • thinning ͍ͨ͠৔߹͸ LoG ϑΟϧλΛ͔͚ͯθϩަࠩ఺Λٻ ΊΔ • Χϥʔը૾ͷΤοδݕग़͸໌౓ɾ৭ɾςΫενϟͳͲͷ౷ܭ ྔ͕༗ޮ • thinning ͞ΕͨΤοδͷ࿈݁͸νΣΠϯίʔυ΍ arc-length parameterization ͕༗ޮ • arc-length parameterization ޙ͸Ϛονϯά΍ϊΠζআڈΛ͠ ΍͍͢ 31

Slide 34

Slide 34 text

References I [1] R. Rao. Image sampling, pyramids, and edge detection. https://courses.cs.washington.edu/courses/cse455/ 09wi/Lects/lect3.pdf, 2009.