7) FOE (Maxmum A Posteror : MAP) MAP 2. ( ) F (1) 2) p(f ; θ) = 1 Z all clques = 1 Z exp [ φ(f ; θ) all clques λ(f ; θ) ] (1) F f φ( ) λ( ) θ Z

Size: px
Start display at page:

Download "7) FOE (Maxmum A Posteror : MAP) MAP 2. ( ) F (1) 2) p(f ; θ) = 1 Z all clques = 1 Z exp [ φ(f ; θ) all clques λ(f ; θ) ] (1) F f φ( ) λ( ) θ Z"

Transcription

1 1 2 2 (MRF) Natural Image pror model usng adaptve mult-varate Gaussan dstrbuton Ketaro Yamauch, 1 Masayuk Tanaka 2 and Masatosh Okutom 2 Many pror models are proposed whch model a dgtal mage by a Markov Random Feld (MRF). However, n many applcatons, only homogeneous MRF whch has the potental functon wth fxed parameters s dscussed. Ths paper proposes a natural mage pror model whch uses non-homogeneous MRF whose parameters are adaptvely desgned. In the proposed model, an mage s decomposed nto low- and hgh-frequency components. Then the hgh-frequency component s modeled by the non-homogeneous MRF. The parameters of the potental functon are adaptvely desgned based on the low-frequency component assocated to the hgh frequency component. In ths paper, we use a multvarate Gaussan functon for the potental functon on the MRF. Fnally, we use the proposed model for denosng by Maxmum a-posteror estmaton. 1. (MRF) ( ) 6) 8),10) 4) 12) 1) 2 8) Roth Felds of Experts (FOE) 6) FOE Roth Student-t Student-t 11) Wess FOE 10) FOE Tappen 8) Tanaka 1 Department of Control and Systems Engneerng, Tokyo Insttute of Technology 2 Graduate School of Scence and Technology, Tokyo Insttute of Technology 1 c 2010 Informaton Processng Socety of Japan

2 7) FOE (Maxmum A Posteror : MAP) MAP 2. ( ) F (1) 2) p(f ; θ) = 1 Z all clques = 1 Z exp [ φ(f ; θ) all clques λ(f ; θ) ] (1) F f φ( ) λ( ) θ Z FOE (2) p(f ; Θ) = 1 Z all clques = 1 Z exp [ φ(f ; θ ) all clques λ(f ; θ ) ] (2) θ Θ = {θ } θ (3) F F L F H F =F L + F H F L = L F (3) F H = F F L L F SN SN (mage salency) 9) (4) 2 c 2010 Informaton Processng Socety of Japan

3 p(f H F L ) = 1 Z all clques φ ( f H ; θ(f L ) ) (4) f H, f L f H φ θ(f L ) f L 3.2 (4) (5) 1 φ ( f H θ(f L ) ) = N ( f H ; µ( f L ), Σ( f L ) ) (5) N ( f H ; µ, Σ ) f H µ Σ f L f H x x f L y y f L f L f L 3.3 f L θ f L f L 1 x y {(f L, f H )} {( f L, f H )} {( f L, f H )} (k, l) D k,l k l (k, l) θ k,l ˆθ k,l ˆθ k,l = arg max φ(v; θ) (6) θ v D k,l θ θ k,l = (µ k,l, Σ k,l ) (7) µ k,l Σ k,l D k,l ˆµ k,l = 1 v (8) D k,l v D k,l ˆΣ k,l = 1 (v ˆµ k,l ) T (v ˆµ k,l ) D k,l v D k,l = 1 (9) v T v ˆµ T k,l ˆµ k,l D k,l v D k,l D k,l D k,l ˆθ k,l f L µ k,l Σ k,l f L 3 c 2010 Informaton Processng Socety of Japan

4 MAP 4.1 G K K K SN K 1 (5) MAP F H (11) p(f H G H, G L ) 1 Z p(gh F H )p(f H ˆF L ) (11) G H, G L G (4),(5) Z F H ˆF L 1 ˆF LL ˆF LH µ k,l Σ k,l 3.4 SN (4) F L = F LL + F LH (10) {F L, F H } 1 {F LL, F LH } 2 K K 2 ˆF L = ˆF LL + ˆF LH (12) (11) 4.2 F H ˆF H ˆF H = arg max p(g H F H )p(f H ˆF L ) (13) F H (13) [ all clques ˆF H 1 = arg mn (g H F H f H ) 2 + ( f H µ( f L ) ) T Σ( f L ) ( 1 f H µ( f L ) ) ] σ 2 N g H, f H G H, F H ˆF H (14) (14) 4 c 2010 Informaton Processng Socety of Japan

5 p(f H, G L ) = p(f H G L )p(g L ) (16) p(f H G L ) = p(f H F L )p(f L G L )df L (17) F H G L G H p(g H F H, G L ) = p(g H F H ) (18) (16) (17) (18) (15) p(f H G H, G L ) = p(gl ) p(g L, G H ) p(gh F H ) p(f H F L )p(f L G L )df L (19) (19) F L (19) p(f L G L ) (20) p(f L G L ) δ(f L ˆF L ) (20) ˆF L = ˆF LL + ˆF LH (21) (11) p(f H G H, G L ) (15) p(f H G H, G L ) = p(gh F H, G L )p(f H, G L ) p(g L, G H ) (15) (20) (19) (23) p(f H G H, G L ) = p(gl ) p(g L, G H ) p(gh F H ) p(f H F L )δ(f L ˆF L )df L (22) (11) 1 Z p(gh F H )p(f H ˆF L ) (23) (21) ˆF L ˆF LL ˆF LH (12) (13) ˆF LL = ˆF LLL + ˆF LLH (24) ˆF LH = arg max F LH p(g LH F LH )p(f LH ˆF LL ) (25) 5 c 2010 Informaton Processng Socety of Japan

6 情報処理学会研究報告 4.3 デノイジング実験 テスト画像に対して標準偏差 20 のガウシアンノイズを人工的に付加し デノイジングを 行った パッチを 3x3 の矩形領域としてモデル化を行い 階層数 K は 1 階層と 5 階層で 実験を行った ここで階層数を K = 5 に選んだのは 経験的に低周波成分のノイズ成分が lena house barbara 十分に小さくなるという理由からである 画像の分解に用いるローパスフィルタ L として boat peppers 図 4 デノイジング実験に用いた画像 標準偏差が 0.6 のガウシアンフィルタを用いた 実験に用いる画像として 図 4 に示す画像 を利用した これらは画像処理のベンチマークとして頻繁に用いられる画像である 表1 処理結果と原画像との類似度の指標として Peak Sgnal Nose Rato (PSNR) を用いる ことにした PSNR は 処理結果と原画像の間の Root Mean Square Error (RMSE) を σe PSNR = 20 log σe 提案法 (K = 5 階層) lena barbara house boat peppers FoE6) Portlla5) et al BM3D3) 提案法 (K = 1 階層) とすると 次の式により計算できる (26) PSNR が大きいほど 原画像との類似度が高くデノイジング性能が優れている 各手法によるデノイジング結果の PSNR[dB] による比較 第 1 階層だけを利用した場合と第 5 階層まで利用した場合の処理結果の比較を 図 5 に示 す また FOE モデルを MAP 推定に用いたデノイジング6) Portlla らの手法5) BM3D ルを MAP 推定に応用しデノイジングを行った テクスチャがぼけてしまうことや 高度な 法3) との PSNR による比較を 表 1 に示す 既存手法に比べ PSNR が劣っているという課題があるものの デノイジングに関して十分 処理結果をみると 第 1 階層だけを利用した場合はノイズの低周波成分が残留しているが な効果が得られた これらの結果により 提案モデルの事前確率分布モデルとしての有効性 第 5 階層まで利用した場合はノイズがほとんど取り除かれていることが分かる barbara を確認した を除いた 4 つの画像では 第 5 階層まで利用した方が PSNR が高い 参 しかし 提案法によるデノイジングでは 階層数を多く利用するほどテクスチャがぼけて しまうことが確認された テクスチャが豊富な barbara では 第 1 階層よりも第 5 階層の 考 文 献 1) Banarjee, J., M.Namboodr, A. and C.V.Jawahar: Contextual Restoraton of Severely Degraded Document Images, IEEE conference on Computer Vson and Pattern Recognton (2009). 2) Bshop, C.M.: Pattern Recognton and Machne Learnng : Informaton Scence and Statstces, Sprnger (2006). 3) Dabov, K., Fo, A., Katkovnk, V. and Egazaran, K.: BM3D Image Denosng wth Shape-Adaptve Prncpal Component Analyss, Workshop on Sgnal Processng Adaptve Sparse Structed Representatons (SPARS) (2009). 4) Freeman, W.T., Jones, T.R. and EgonC, P.: Example-Based Super-Resoluton, Computer Graphcs and Applcaton, Vol.22 (2002). 5) Portlla, J., Strela, V., J.Wanwrght, M. and P.Smoncell, E.: Image Denosng usng Scale Mxture of Gaussans n the Wavelet Doman, IEEE Transactons on Image Processng, Vol.12, No.11 (2003). 処理結果の方が服の縞模様が失われているため 結果として PSNR が低い また Portlla らの手法や BM3D 法のようなデノイジングに特化した高性能な手法に比 べると 提案モデルを利用したデノイジングでは PSNR が低い しかし 提案モデルを単 純に応用したにも関わらず FOE モデルによるデノイジングに近い性能を有している こ れらの結果は提案モデルの事前確率分布モデルとしての有効性を示していると考える 5. む す び 本論文では 画像の高周波成分を非均質マルコフ確率場によってモデル化する方法を提案 した 提案モデルでは 非均質マルコフ確率場のパラメータは対応する低周波成分に応じて 変化する 本論文では ポテンシャル関数が多変量正規分布であるものを考えた 提案モデ 6 c 2010 Informaton Processng Socety of Japan

7 情報処理学会研究報告 (a) 原画像 (b) ノイズを合成した画像 (c) 第 1 階層だけを利用 図5 (d) 第 5 階層までを利用 (e)foe モデルによる結果6) (f)bm3d 法による結果3) デノイジング処理結果の比較 (lena, barbara, house) Vson and Pattern Recognton (2007). 9) Valent, R., Sebe, N. and Gevers, T.: Image Salency by Isocentrc Curvedness and Color, IEEE Conference on Computer Vson (2009). 10) Wess, Y. and T.Freeman, W.: What makes a good model of natural mages?, Computer Vson and Pattern Recognton (2007). 11) Wellng, M., Hnton, G. and Osndero, S.: Learnng Sparse Topographc Representatons wth Products of Student-t Dstrbutons, NIPS, Vol.15 (2003). 6) S.Roth and M.J.Black: Felds of Experts: A Framework for Learnng Image Prors., Computer Vson and Pattern Recognton, Vol.2 (2005). 7) Tanaka, M. and Okutom, M.: Locally Adaptve Learnng for Translaton-Varant MRF Image Prors, IEEE conference on Computer Vson and Pattern Recognton (2008). 8) Tappen, M. F., Lu, C., Adelson, E. H. and Freeman, W. T.: Learnng Gaussan Condtonal Random Felds for Low-Level Vson, IEEE conference on Computer 7 c 2010 Informaton Processng Socety of Japan

8 12) Woodford, O.J., Rother, C. and Kolmogorov, V.: A Grobal Perspectve on MAP Inference for Low-Level Vson, ICCV (2009). 8 c 2010 Informaton Processng Socety of Japan

企業情報 / 研究開発体制 | Ricoh Japan

企業情報 / 研究開発体制 | Ricoh Japan 画 像 のε 近 傍 に 基 づくカラーCCDノイズ 低 減 Nose Reducton for Color CCD Image Sensors Based on ε-neghborhood of Images * 原 崇 之 Takayuk HARA ** 関 海 克 Hake GUAN 要 旨 デジタルカメラやイメージスキャナで 使 用 されるCCDカラーイメージセンサで 撮 影 された 画 像

More information

ohpmain.dvi

ohpmain.dvi fujisawa@ism.ac.jp 1 Contents 1. 2. 3. 4. γ- 2 1. 3 10 5.6, 5.7, 5.4, 5.5, 5.8, 5.5, 5.3, 5.6, 5.4, 5.2. 5.5 5.6 +5.7 +5.4 +5.5 +5.8 +5.5 +5.3 +5.6 +5.4 +5.2 =5.5. 10 outlier 5 5.6, 5.7, 5.4, 5.5, 5.8,

More information

4. C i k = 2 k-means C 1 i, C 2 i 5. C i x i p [ f(θ i ; x) = (2π) p 2 Vi 1 2 exp (x µ ] i) t V 1 i (x µ i ) 2 BIC BIC = 2 log L( ˆθ i ; x i C i ) + q

4. C i k = 2 k-means C 1 i, C 2 i 5. C i x i p [ f(θ i ; x) = (2π) p 2 Vi 1 2 exp (x µ ] i) t V 1 i (x µ i ) 2 BIC BIC = 2 log L( ˆθ i ; x i C i ) + q x-means 1 2 2 x-means, x-means k-means Bayesian Information Criterion BIC Watershed x-means Moving Object Extraction Using the Number of Clusters Determined by X-means Clustering Naoki Kubo, 1 Kousuke

More information

main.dvi

main.dvi THE INSTITUTE OF ELECTRONICS, INFORMATION AND COMMUNICATION ENGINEERS TECHNICAL REPORT OF IEICE. ( ) Estmaton and Analyss of Topc Models n Tme Seres Japanese / Chnese News and Blogs Shuo HU,LyZHENG, Yusuke

More information

> σ, σ j, j σ j, σ j j σ σ j σ j (t) = σ (t ) σ j (t) = σ () j(t ) n j σ, σ j R lm σ = σ j, j V (8) t σ R σ d R lm σ = σ d V (9) t Fg.. Communcaton ln

> σ, σ j, j σ j, σ j j σ σ j σ j (t) = σ (t ) σ j (t) = σ () j(t ) n j σ, σ j R lm σ = σ j, j V (8) t σ R σ d R lm σ = σ d V (9) t Fg.. Communcaton ln IIC-- Dstrbuted Cooperatve Atttude Control for Multple Rgd Bodes wth Communcaton Delay Yoshhro achbana, oru Namerkawa (Keo Unversty) Abstract hs paper descrbes dstrbuted cooperatve atttude consensus and

More information

IPSJ SIG Technical Report Vol.2013-MUS-100 No /9/1 1,a) [1 9] 1 3 MIDI MIDI (Interonset interval) 1 NTT NTT Communication Science Labo

IPSJ SIG Technical Report Vol.2013-MUS-100 No /9/1 1,a) [1 9] 1 3 MIDI MIDI (Interonset interval) 1 NTT NTT Communication Science Labo 1,a) 1 1 1. [1 9] 1 3 MIDI MIDI (Interonset nterval) 1 N N Communcaton Scence Laboratores, Atsug, Kanagawa 43 0198, Japan he Insttute of Statstcal Mathematcs, achkawa, okyo 190 856 a) ohsh.yasunor@lab.ntt.co.jp

More information

& 3 3 ' ' (., (Pixel), (Light Intensity) (Random Variable). (Joint Probability). V., V = {,,, V }. i x i x = (x, x,, x V ) T. x i i (State Variable),

& 3 3 ' ' (., (Pixel), (Light Intensity) (Random Variable). (Joint Probability). V., V = {,,, V }. i x i x = (x, x,, x V ) T. x i i (State Variable), .... Deeping and Expansion of Large-Scale Random Fields and Probabilistic Image Processing Kazuyuki Tanaka The mathematical frameworks of probabilistic image processing are formulated by means of Markov

More information

3 1, 1, 1, 1 3D Environment Measurement Using Binocular Stereo and Motion Stereo by Mobile Robot with Omnidirectional Stereo Camera Shinichi GOTO 1, R

3 1, 1, 1, 1 3D Environment Measurement Using Binocular Stereo and Motion Stereo by Mobile Robot with Omnidirectional Stereo Camera Shinichi GOTO 1, R 3 1, 1, 1, 1 3D Envronment Measurement Usng Bnocular Stereo and Moton Stereo by Moble Robot wth Omndrectonal Stereo Camera Shnch GOTO 1, Ryosuke KAWANISHI 1, Atsush YAMASHITA 1 and Toru KANEKO 1 1 Department

More information

IPSJ SIG Techncal Report 2. RangeBased RangeFree. 2.1 Rangebased RangeBased TDOA(Tme Dfference Of Arrval) TOA(Tme Of Arrval) TDOA TDOA Actve Bat 2) Cr

IPSJ SIG Techncal Report 2. RangeBased RangeFree. 2.1 Rangebased RangeBased TDOA(Tme Dfference Of Arrval) TOA(Tme Of Arrval) TDOA TDOA Actve Bat 2) Cr IPSJ SIG Techncal Report 1 2 2 (SOM, Self Organzng Maps) 7). Self-Organzng Localzaton for Wreless Sensor Networks on Ansotropc Topology Yuto Takashma, 1 Naotosh Adach 2 and Yasuhsa Takzawa 2 On wreless

More information

2 Poisson Image Editing DC DC 2 Poisson Image Editing Agarwala 3 4 Agarwala Poisson Image Editing Poisson Image Editing f(u) u 2 u = (x

2 Poisson Image Editing DC DC 2 Poisson Image Editing Agarwala 3 4 Agarwala Poisson Image Editing Poisson Image Editing f(u) u 2 u = (x 1 Poisson Image Editing Poisson Image Editing Stabilization of Poisson Equation for Gradient-Based Image Composing Ryo Kamio Masayuki Tanaka Masatoshi Okutomi Poisson Image Editing is the image composing

More information

10_4.dvi

10_4.dvi Vol.44, No.1, 1/7 28 Synchronzed Control for Blateral Teleoperaton wth Dfferent Confguratons and Communcaton Delays Hsanosuke Kawada,KoueYoshda and Toru Namerkawa Ths paper addresses the problem of the

More information

N cos s s cos ψ e e e e 3 3 e e 3 e 3 e

N cos s s cos ψ e e e e 3 3 e e 3 e 3 e 3 3 5 5 5 3 3 7 5 33 5 33 9 5 8 > e > f U f U u u > u ue u e u ue u ue u e u e u u e u u e u N cos s s cos ψ e e e e 3 3 e e 3 e 3 e 3 > A A > A E A f A A f A [ ] f A A e > > A e[ ] > f A E A < < f ; >

More information

JGSS統計分析セミナー2009-傾向スコアを用いた因果分析-

JGSS統計分析セミナー2009-傾向スコアを用いた因果分析- 日本版総合的社会調査共同研究拠点研究論文集 [10] JGSS Research Seres No.7 JGSS 2009 JGSS JGSS Statstcal Analyss Semnar: Causalty Analyss based on the Propensty Score Kana MIWA JGSS Research Center Osaka Unversty of Commerce

More information

画像処理工学

画像処理工学 画像処理工学 画像の空間周波数解析とテクスチャ特徴 フーリエ変換の基本概念 信号波形のフーリエ変換 信号波形を周波数の異なる三角関数 ( 正弦波など ) に分解する 逆に, 周波数の異なる三角関数を重ねあわせることにより, 任意の信号波形を合成できる 正弦波の重ね合わせによる矩形波の表現 フーリエ変換の基本概念 フーリエ変換 次元信号 f (t) のフーリエ変換 変換 ( ω) ( ) ωt F f

More information

(MIRU2008) HOG Histograms of Oriented Gradients (HOG)

(MIRU2008) HOG Histograms of Oriented Gradients (HOG) (MIRU2008) 2008 7 HOG - - E-mail: katsu0920@me.cs.scitec.kobe-u.ac.jp, {takigu,ariki}@kobe-u.ac.jp Histograms of Oriented Gradients (HOG) HOG Shape Contexts HOG 5.5 Histograms of Oriented Gradients D Human

More information

IPSJ SIG Technical Report Vol.2010-CVIM-172 No /5/ Object Tracking Based on Generative Appearance Model 1. ( 1 ) ( 2 ) ( 3 ) 1 3) T

IPSJ SIG Technical Report Vol.2010-CVIM-172 No /5/ Object Tracking Based on Generative Appearance Model 1. ( 1 ) ( 2 ) ( 3 ) 1 3) T 1 2 2 3 1 Objec Tracking Based on Generaive Appearance Model 1. ( 1 ) ( 2 ) ( 3 ) 1 3) Tasuya YONEKAWA, 1 Kazuhiko KAWAMOTO, 2 Asushi IMIYA 2 and Akihiro SUGIMOTO 3 We propose a mehod for racking objecs

More information

Microsoft Word - 田口君最終報告(修正後).doc

Microsoft Word - 田口君最終報告(修正後).doc 2005 7 1 2005 7 2 1 2 2 4 Data Envelopment Analyss 4 Stochastc Fronter Analyss 5 Non-mnmum Cost Functon 6 7 8 9 9 9 13 13 13 14 14 17 19 20 21 24 24 24 25 25 28 31 36 37 40 3 2 2005 7 1950 1960 2003 1004

More information

Microsoft Word - Ï. ÏÌáÉ + íÇÓãíä.doc

Microsoft Word - Ï. ÏÌáÉ + íÇÓãíä.doc 3948 98789 98963974 9897994 9498876 046 06 /6 / 50 86 086 0 6749 798 98945799 988997894 96799 74 98599446 988997894 9894 9.6 9.6 4994 77 98959 9597 748 99949 7.6 9.6 94.6 989683 9 54 89 9853 799 98599446

More information

橡jttc2.PDF

橡jttc2.PDF 1 ( ) 1 GA GA GA MOGA (Multple-Objectve Genetc Algorthm) GA GA GA MOGA GA GA MOGA GA GA 3.1MOGA ( ) x x j f = f, f, 1 2 L, f q x x j x j f ( x ) f ( x ) f ( x ) f ( x ) L f ( x ) f ( x ) ( ) ( ) 1 1 j

More information

IPSJ SIG Technical Report Vol.2010-CVIM-170 No /1/ Visual Recognition of Wire Harnesses for Automated Wiring Masaki Yoneda, 1 Ta

IPSJ SIG Technical Report Vol.2010-CVIM-170 No /1/ Visual Recognition of Wire Harnesses for Automated Wiring Masaki Yoneda, 1 Ta 1 1 1 1 2 1. Visual Recognition of Wire Harnesses for Automated Wiring Masaki Yoneda, 1 Takayuki Okatani 1 and Koichiro Deguchi 1 This paper presents a method for recognizing the pose of a wire harness

More information

4 4 2 RAW 4 4 4 (PCA) 4 4 4 4 RAW RAW [5] 4 RAW 4 Park [12] Park 2 RAW RAW 2 RAW y = Mx + n. (1) y RAW x RGB M CFA n.. R G B σr 2, σ2 G, σ2 B D n ( )

4 4 2 RAW 4 4 4 (PCA) 4 4 4 4 RAW RAW [5] 4 RAW 4 Park [12] Park 2 RAW RAW 2 RAW y = Mx + n. (1) y RAW x RGB M CFA n.. R G B σr 2, σ2 G, σ2 B D n ( ) RAW 4 E-mail: hakiyama@ok.ctrl.titech.ac.jp Abstract RAW RAW RAW RAW RAW 4 RAW RAW RAW 1 (CFA) CFA Bayer CFA [1] RAW CFA 1 2 [2, 3, 4, 5]. RAW RAW RAW RAW 3 [2, 3, 4, 5] (AWGN) [13, 14] RAW 2 RAW RAW RAW

More information

201711grade1ouyou.pdf

201711grade1ouyou.pdf 2017 11 26 1 2 52 3 12 13 22 23 32 33 42 3 5 3 4 90 5 6 A 1 2 Web Web 3 4 1 2... 5 6 7 7 44 8 9 1 2 3 1 p p >2 2 A 1 2 0.6 0.4 0.52... (a) 0.6 0.4...... B 1 2 0.8-0.2 0.52..... (b) 0.6 0.52.... 1 A B 2

More information

THE INSTITUTE OF ELECTRONICS, INFORMATION AND COMMUNICATION ENGINEERS TECHNICAL REPORT OF IEICE. y y y y {3{1

THE INSTITUTE OF ELECTRONICS, INFORMATION AND COMMUNICATION ENGINEERS TECHNICAL REPORT OF IEICE. y y y y {3{1 THE INSTITUTE OF ELECTRONICS, INFORMATION AND COMMUNICATION ENGINEERS TECHNICAL REPORT OF IEICE. y y y y 3-8656 7{3{ E-mal: ynamura@jsk.t.u-tokyo.ac.jp, yyftane,nakamurag@ynl.t.u-tokyo.ac.jp Proto-Symbol

More information

集中理論談話会 #9 Bhat, C.R., Sidharthan, R.: A simulation evaluation of the maximum approximate composite marginal likelihood (MACML) estimator for mixed mu

集中理論談話会 #9 Bhat, C.R., Sidharthan, R.: A simulation evaluation of the maximum approximate composite marginal likelihood (MACML) estimator for mixed mu 集中理論談話会 #9 Bhat, C.R., Sidharthan, R.: A simulation evaluation of the maximum approximate composite marginal likelihood (MACML) estimator for mixed multinomial probit models, Transportation Research Part

More information

Vol. 44 No. SIG 9(CVIM 7) ) 2) 1) 1 2) 3 7) 1) 2) 3 3) 4) 5) (a) (d) (g) (b) (e) (h) No Convergence? End (f) (c) Yes * ** * ** 1

Vol. 44 No. SIG 9(CVIM 7) ) 2) 1) 1 2) 3 7) 1) 2) 3 3) 4) 5) (a) (d) (g) (b) (e) (h) No Convergence? End (f) (c) Yes * ** * ** 1 Vol. 44 No. SIG 9(CVIM 7) July 2003, Robby T. Tan, 1 Estimating Illumination Position, Color and Surface Reflectance Properties from a Single Image Kenji Hara,, Robby T. Tan, Ko Nishino, Atsushi Nakazawa,

More information

258 5) GPS 1 GPS 6) GPS DP 7) 8) 10) GPS GPS 2 3 4 5 2. 2.1 3 1) GPS Global Positioning System

258 5) GPS 1 GPS 6) GPS DP 7) 8) 10) GPS GPS 2 3 4 5 2. 2.1 3 1) GPS Global Positioning System Vol. 52 No. 1 257 268 (Jan. 2011) 1 2, 1 1 measurement. In this paper, a dynamic road map making system is proposed. The proposition system uses probe-cars which has an in-vehicle camera and a GPS receiver.

More information

? (EM),, EM? (, 2004/ 2002) von Mises-Fisher ( 2004) HMM (MacKay 1997) LDA (Blei et al. 2001) PCFG ( 2004)... Variational Bayesian methods for Natural

? (EM),, EM? (, 2004/ 2002) von Mises-Fisher ( 2004) HMM (MacKay 1997) LDA (Blei et al. 2001) PCFG ( 2004)... Variational Bayesian methods for Natural SLC Internal tutorial Daichi Mochihashi daichi.mochihashi@atr.jp ATR SLC 2005.6.21 (Tue) 13:15 15:00@Meeting Room 1 Variational Bayesian methods for Natural Language Processing p.1/30 ? (EM),, EM? (, 2004/

More information

THE INSTITUTE OF ELECTRONICS, INFORMATION AND COMMUNICATION ENGINEERS TECHNICAL REPORT OF IEICE.

THE INSTITUTE OF ELECTRONICS, INFORMATION AND COMMUNICATION ENGINEERS TECHNICAL REPORT OF IEICE. THE INSTITUTE OF ELECTRONICS, INFORMATION AND COMMUNICATION ENGINEERS TECHNICAL REPORT OF IEICE. E-mail: {ytamura,takai,tkato,tm}@vision.kuee.kyoto-u.ac.jp Abstract Current Wave Pattern Analysis for Anomaly

More information

Vol.0 No.0, 2002 : Facial Expression Representation based on Timing Structures in Faces Takatsugu Hirayama 1, Hiroaki Kawashima 1, Masahiro Ni

Vol.0 No.0, 2002 : Facial Expression Representation based on Timing Structures in Faces Takatsugu Hirayama 1, Hiroaki Kawashima 1, Masahiro Ni Vol. No., 22 : 1 1 1 1 Facal Expresson Representaton based on Tmng Structures n Faces Taatsugu Hrayama 1, Hroa Kawashma 1, Masahro Nshyama 1, and Taash Matsuyama 1 Abstract Ths paper presents a method

More information

通勤混雑と家賃関数*

通勤混雑と家賃関数* CIRJE-J-30 CIRJE 000 8 5 9 3 Estmaton of Fatgue Cost of Commutng Congeston and Optmal Congeston Fare Ths paper has three ams. Frst, we estmate a hedonc housng rent functon along the Chuo Lne n Tokyo wth

More information

1 Fig. 1 Extraction of motion,.,,, 4,,, 3., 1, 2. 2.,. CHLAC,. 2.1,. (256 ).,., CHLAC. CHLAC, HLAC. 2.3 (HLAC ) r,.,. HLAC. N. 2 HLAC Fig. 2

1 Fig. 1 Extraction of motion,.,,, 4,,, 3., 1, 2. 2.,. CHLAC,. 2.1,. (256 ).,., CHLAC. CHLAC, HLAC. 2.3 (HLAC ) r,.,. HLAC. N. 2 HLAC Fig. 2 CHLAC 1 2 3 3,. (CHLAC), 1).,.,, CHLAC,.,. Suspicious Behavior Detection based on CHLAC Method Hideaki Imanishi, 1 Toyohiro Hayashi, 2 Shuichi Enokida 3 and Toshiaki Ejima 3 We have proposed a method for

More information

3 2 2 (1) (2) (3) (4) 4 4 AdaBoost 2. [11] Onishi&Yoda [8] Iwashita&Stoica [5] 4 [3] 3. 3 (1) (2) (3)

3 2 2 (1) (2) (3) (4) 4 4 AdaBoost 2. [11] Onishi&Yoda [8] Iwashita&Stoica [5] 4 [3] 3. 3 (1) (2) (3) (MIRU2012) 2012 8 820-8502 680-4 E-mail: {d kouno,shimada,endo}@pluto.ai.kyutech.ac.jp (1) (2) (3) (4) 4 AdaBoost 1. Kanade [6] CLAFIC [12] EigenFace [10] 1 1 2 1 [7] 3 2 2 (1) (2) (3) (4) 4 4 AdaBoost

More information

IPSJ SIG Technical Report 1,a) 1,b) 1,c) 1,d) 2,e) 2,f) 2,g) 1. [1] [2] 2 [3] Osaka Prefecture University 1 1, Gakuencho, Naka, Sakai,

IPSJ SIG Technical Report 1,a) 1,b) 1,c) 1,d) 2,e) 2,f) 2,g) 1. [1] [2] 2 [3] Osaka Prefecture University 1 1, Gakuencho, Naka, Sakai, 1,a) 1,b) 1,c) 1,d) 2,e) 2,f) 2,g) 1. [1] [2] 2 [3] 1 599 8531 1 1 Osaka Prefecture University 1 1, Gakuencho, Naka, Sakai, Osaka 599 8531, Japan 2 565 0871 Osaka University 1 1, Yamadaoka, Suita, Osaka

More information

Part () () Γ Part ,

Part () () Γ Part , Contents a 6 6 6 6 6 6 6 7 7. 8.. 8.. 8.3. 8 Part. 9. 9.. 9.. 3. 3.. 3.. 3 4. 5 4.. 5 4.. 9 4.3. 3 Part. 6 5. () 6 5.. () 7 5.. 9 5.3. Γ 3 6. 3 6.. 3 6.. 3 6.3. 33 Part 3. 34 7. 34 7.. 34 7.. 34 8. 35

More information

untitled

untitled 17 5 13 1 2 1.1... 2 1.2... 2 1.3... 3 2 3 2.1... 3 2.2... 5 3 6 3.1... 6 3.2... 7 3.3 t... 7 3.4 BC a... 9 3.5... 10 4 11 1 1 θ n ˆθ. ˆθ, ˆθ, ˆθ.,, ˆθ.,.,,,. 1.1 ˆθ σ 2 = E(ˆθ E ˆθ) 2 b = E(ˆθ θ). Y 1,,Y

More information

COE-RES Discussion Paper Series Center of Excellence Project The Normative Evaluation and Social Choice of Contemporary Economic Systems Graduate Scho

COE-RES Discussion Paper Series Center of Excellence Project The Normative Evaluation and Social Choice of Contemporary Economic Systems Graduate Scho COE-RES Discussion Paper Series Center of Excellence Project The Normative Evaluation and Social Choice of Contemporary Economic Systems Graduate School of Economics and Institute of Economic Research

More information

xx/xx Vol. Jxx A No. xx 1 Fig. 1 PAL(Panoramic Annular Lens) PAL(Panoramic Annular Lens) PAL (2) PAL PAL 2 PAL 3 2 PAL 1 PAL 3 PAL PAL 2. 1 PAL

xx/xx Vol. Jxx A No. xx 1 Fig. 1 PAL(Panoramic Annular Lens) PAL(Panoramic Annular Lens) PAL (2) PAL PAL 2 PAL 3 2 PAL 1 PAL 3 PAL PAL 2. 1 PAL PAL On the Precision of 3D Measurement by Stereo PAL Images Hiroyuki HASE,HirofumiKAWAI,FrankEKPAR, Masaaki YONEDA,andJien KATO PAL 3 PAL Panoramic Annular Lens 1985 Greguss PAL 1 PAL PAL 2 3 2 PAL DP

More information

研究シリーズ第40号

研究シリーズ第40号 165 PEN WPI CPI WAGE IIP Feige and Pearce 166 167 168 169 Vector Autoregression n (z) z z p p p zt = φ1zt 1 + φ2zt 2 + + φ pzt p + t Cov( 0 ε t, ε t j )= Σ for for j 0 j = 0 Cov( ε t, zt j ) = 0 j = >

More information

情報処理学会研究報告 IPSJ SIG Technical Report 電子透かしに平均値攻撃耐性を持たせるための一手法 久世慎吾 岩村惠市 結託攻撃のひとつとして多くのユーザーが結託していくつかの透かし画像を平均化することで 透かし情報などを打ち消す平均値攻撃がある 画像圧縮やノイズ付加などの通

情報処理学会研究報告 IPSJ SIG Technical Report 電子透かしに平均値攻撃耐性を持たせるための一手法 久世慎吾 岩村惠市 結託攻撃のひとつとして多くのユーザーが結託していくつかの透かし画像を平均化することで 透かし情報などを打ち消す平均値攻撃がある 画像圧縮やノイズ付加などの通 電子透かしに平均値攻撃耐性を持たせるための一手法 久世慎吾 岩村惠市 結託攻撃のひとつとして多くのユーザーが結託していくつかの透かし画像を平均化することで 透かし情報などを打ち消す平均値攻撃がある 画像圧縮やノイズ付加などの通常の電子透かし画像に対する攻撃は画質を劣化させるが 平均値攻撃は多くの画像が集まれば集まるほど 平均化された画像を原画像に近づけ画質を向上させる よって 平均値攻撃は電子透かしにとって最も脅威とすべき攻撃と考えられるが

More information

(a) 1 (b) 3. Gilbert Pernicka[2] Treibitz Schechner[3] Narasimhan [4] Kim [5] Nayar [6] [7][8][9] 2. X X X [10] [11] L L t L s L = L t + L s

(a) 1 (b) 3. Gilbert Pernicka[2] Treibitz Schechner[3] Narasimhan [4] Kim [5] Nayar [6] [7][8][9] 2. X X X [10] [11] L L t L s L = L t + L s 1 1 1, Extraction of Transmitted Light using Parallel High-frequency Illumination Kenichiro Tanaka 1 Yasuhiro Mukaigawa 1 Yasushi Yagi 1 Abstract: We propose a new sharpening method of transmitted scene

More information

IPSJ SIG Technical Report GPS LAN GPS LAN GPS LAN Location Identification by sphere image and hybrid sensing Takayuki Katahira, 1 Yoshio Iwai 1

IPSJ SIG Technical Report GPS LAN GPS LAN GPS LAN Location Identification by sphere image and hybrid sensing Takayuki Katahira, 1 Yoshio Iwai 1 1 1 1 GPS LAN GPS LAN GPS LAN Location Identification by sphere image and hybrid sensing Takayuki Katahira, 1 Yoshio Iwai 1 and Hiroshi Ishiguro 1 Self-location is very informative for wearable systems.

More information

Microsoft PowerPoint - SSII_harada pptx

Microsoft PowerPoint - SSII_harada pptx The state of the world The gathered data The processed data w d r I( W; D) I( W; R) The data processing theorem states that data processing can only destroy information. David J.C. MacKay. Information

More information

wakate2005-tobari-ver2.ppt

wakate2005-tobari-ver2.ppt 8 005 3 17 Outlne 1. Introducton. 3. 4. Introducton Image of MUSES-C on engne JAXA b = 4ε 9 q m V d 3 a large-scale) MHD MPD chokng VASIMR Super-Alfvénc flow plasma detachment HITOP(HIgh densty TOhoku

More information

28 Horizontal angle correction using straight line detection in an equirectangular image

28 Horizontal angle correction using straight line detection in an equirectangular image 28 Horizontal angle correction using straight line detection in an equirectangular image 1170283 2017 3 1 2 i Abstract Horizontal angle correction using straight line detection in an equirectangular image

More information

2.2 RPS Feed-n Tarff ECO [3] ECO 1 ECO ECO 1 DC ECO 2.3 [9] ECO ECO ECO ECO Natural Actor-Crtc 3 3.1 ECO ECO (Electrc Power Cluster Orented Network) [

2.2 RPS Feed-n Tarff ECO [3] ECO 1 ECO ECO 1 DC ECO 2.3 [9] ECO ECO ECO ECO Natural Actor-Crtc 3 3.1 ECO ECO (Electrc Power Cluster Orented Network) [ Natural Actor-Crtc Desgn of autonomous tradng agent usng natural actor-crtc method to realze a locally produced and consumed electrc energy network Tadahro Tanguch Keta Takag Kazutosh Sakakbara Ikuko Nshkawa

More information

, ( ξ/) ξ(x), ( ξ/) x = x 1,. ξ ξ ( ξ, u) = 0. M LS ξ ξ (6) u,, u M LS 3).,.. ξ x ξ = ξ(x),, 1. J = (ξ ξ, V [ξ ] 1 (ξ ξ )) (7) ( ξ, u) = 0, = 1,..., N

, ( ξ/) ξ(x), ( ξ/) x = x 1,. ξ ξ ( ξ, u) = 0. M LS ξ ξ (6) u,, u M LS 3).,.. ξ x ξ = ξ(x),, 1. J = (ξ ξ, V [ξ ] 1 (ξ ξ )) (7) ( ξ, u) = 0, = 1,..., N 1,,.,.. Maximum Likelihood Estimation for Geometric Fitting Yasuyuki Sugaya 1 Geometric fitting, the problem which estimates a geometric model of a scene from extracted image data, is one of the most fundamental

More information

(1.2) T D = 0 T = D = 30 kn 1.2 (1.4) 2F W = 0 F = W/2 = 300 kn/2 = 150 kn 1.3 (1.9) R = W 1 + W 2 = = 1100 N. (1.9) W 2 b W 1 a = 0

(1.2) T D = 0 T = D = 30 kn 1.2 (1.4) 2F W = 0 F = W/2 = 300 kn/2 = 150 kn 1.3 (1.9) R = W 1 + W 2 = = 1100 N. (1.9) W 2 b W 1 a = 0 1 1 1.1 1.) T D = T = D = kn 1. 1.4) F W = F = W/ = kn/ = 15 kn 1. 1.9) R = W 1 + W = 6 + 5 = 11 N. 1.9) W b W 1 a = a = W /W 1 )b = 5/6) = 5 cm 1.4 AB AC P 1, P x, y x, y y x 1.4.) P sin 6 + P 1 sin 45

More information

TD(0) Q AC (Reward): () Pr(r t+1 s t+1 = s,s t = s, a t = a) t R a ss = E(r t+1 s t+1 = s,s t = s, a t = a) R t = r t+1 + γr t γ T r t+t +1 = T

TD(0) Q AC (Reward): () Pr(r t+1 s t+1 = s,s t = s, a t = a) t R a ss = E(r t+1 s t+1 = s,s t = s, a t = a) R t = r t+1 + γr t γ T r t+t +1 = T () 2009 TD(0) Q AC 2009 1/42 2009 2/42 TD(0) Q AC (Renforcement Learnng) : (polcy) Acton: a t Agent (= Controller) Envronment (= Controlled object) State: s t Reward: r t TD(0) Q AC (Envronment) (Markov

More information

14 2 5

14 2 5 14 2 5 i ii Surface Reconstruction from Point Cloud of Human Body in Arbitrary Postures Isao MORO Abstract We propose a method for surface reconstruction from point cloud of human body in arbitrary postures.

More information

,,.,.,,.,.,.,.,,.,..,,,, i

,,.,.,,.,.,.,.,,.,..,,,, i 22 A person recognition using color information 1110372 2011 2 13 ,,.,.,,.,.,.,.,,.,..,,,, i Abstract A person recognition using color information Tatsumo HOJI Recently, for the purpose of collection of

More information

IPSJ SIG Technical Report Vol.2016-CE-137 No /12/ e β /α α β β / α A judgment method of difficulty of task for a learner using simple

IPSJ SIG Technical Report Vol.2016-CE-137 No /12/ e β /α α β β / α A judgment method of difficulty of task for a learner using simple 1 2 3 4 5 e β /α α β β / α A judgment method of difficulty of task for a learner using simple electroencephalograph Katsuyuki Umezawa 1 Takashi Ishida 2 Tomohiko Saito 3 Makoto Nakazawa 4 Shigeichi Hirasawa

More information

塗装深み感の要因解析

塗装深み感の要因解析 17 Analysis of Factors for Paint Depth Feeling Takashi Wada, Mikiko Kawasumi, Taka-aki Suzuki ( ) ( ) ( ) The appearance and quality of objects are controlled by paint coatings on the surfaces of the objects.

More information

Wavelet HSI / [1] JPEG2000 9/7Wavelet [2][6] 2:1 9/7Wavelet Wavelet 80 Wavelet i

Wavelet HSI / [1] JPEG2000 9/7Wavelet [2][6] 2:1 9/7Wavelet Wavelet 80 Wavelet i 17 Wavelet Image Enhancement by Wavelet Transform 1060326 2006 3 10 Wavelet HSI / [1] JPEG2000 9/7Wavelet [2][6] 2:1 9/7Wavelet Wavelet 80 Wavelet i Abstract Image Enhancement by Wavelet Transform Yuichi

More information

.2 ρ dv dt = ρk grad p + 3 η grad (divv) + η 2 v.3 divh = 0, rote + c H t = 0 dive = ρ, H = 0, E = ρ, roth c E t = c ρv E + H c t = 0 H c E t = c ρv T

.2 ρ dv dt = ρk grad p + 3 η grad (divv) + η 2 v.3 divh = 0, rote + c H t = 0 dive = ρ, H = 0, E = ρ, roth c E t = c ρv E + H c t = 0 H c E t = c ρv T NHK 204 2 0 203 2 24 ( ) 7 00 7 50 203 2 25 ( ) 7 00 7 50 203 2 26 ( ) 7 00 7 50 203 2 27 ( ) 7 00 7 50 I. ( ν R n 2 ) m 2 n m, R = e 2 8πε 0 hca B =.09737 0 7 m ( ν = ) λ a B = 4πε 0ħ 2 m e e 2 = 5.2977

More information

1 Tokyo Daily Rainfall (mm) Days (mm)

1 Tokyo Daily Rainfall (mm) Days (mm) ( ) r-taka@maritime.kobe-u.ac.jp 1 Tokyo Daily Rainfall (mm) 0 100 200 300 0 10000 20000 30000 40000 50000 Days (mm) 1876 1 1 2013 12 31 Tokyo, 1876 Daily Rainfall (mm) 0 50 100 150 0 100 200 300 Tokyo,

More information

it-ken_open.key

it-ken_open.key 深層学習技術の進展 ImageNet Classification 画像認識 音声認識 自然言語処理 機械翻訳 深層学習技術は これらの分野において 特に圧倒的な強みを見せている Figure (Left) Eight ILSVRC-2010 test Deep images and the cited4: from: ``ImageNet Classification with Networks et

More information

IPSJ SIG Technical Report Vol.2015-CVIM-196 No /3/6 1,a) 1,b) 1,c) U,,,, The Camera Position Alignment on a Gimbal Head for Fixed Viewpoint Swi

IPSJ SIG Technical Report Vol.2015-CVIM-196 No /3/6 1,a) 1,b) 1,c) U,,,, The Camera Position Alignment on a Gimbal Head for Fixed Viewpoint Swi 1,a) 1,b) 1,c) U,,,, The Camera Position Alignment on a Gimbal Head for Fixed Viewpoint Swiveling using a Misalignment Model Abstract: When the camera sets on a gimbal head as a fixed-view-point, it is

More information

R = Ar l B r l. A, B A, B.. r 2 R r = r2 [lar r l B r l2 ]=larl l B r l.2 r 2 R = [lar l l Br ] r r r = ll Ar l ll B = ll R rl.3 sin θ Θ = ll.4 Θsinθ

R = Ar l B r l. A, B A, B.. r 2 R r = r2 [lar r l B r l2 ]=larl l B r l.2 r 2 R = [lar l l Br ] r r r = ll Ar l ll B = ll R rl.3 sin θ Θ = ll.4 Θsinθ .3.2 3.3.2 Spherical Coorinates.5: Laplace 2 V = r 2 r 2 x = r cos φ sin θ, y = r sin φ sin θ, z = r cos θ.93 r 2 sin θ sin θ θ θ r 2 sin 2 θ 2 V =.94 2.94 z V φ Laplace r 2 r 2 r 2 sin θ.96.95 V r 2 R

More information

ρ /( ρ) + ( q, v ) : ( q, v ), L < q < q < q < L 0 0 ( t) ( q ( t), v ( t)) dq ( t) v ( t) lmr + 0 Φ( r) dt lmr + 0 Φ ( r) dv ( t) Φ ( q ( t) q ( t)) + Φ ( q+ ( t) q ( t)) dt ( ) < 0 ( q (0), v (0)) (

More information

DPA,, ShareLog 3) 4) 2.2 Strino Strino STRain-based user Interface with tacticle of elastic Natural ObjectsStrino 1 Strino ) PC Log-Log (2007 6)

DPA,, ShareLog 3) 4) 2.2 Strino Strino STRain-based user Interface with tacticle of elastic Natural ObjectsStrino 1 Strino ) PC Log-Log (2007 6) 1 2 1 3 Experimental Evaluation of Convenient Strain Measurement Using a Magnet for Digital Public Art Junghyun Kim, 1 Makoto Iida, 2 Takeshi Naemura 1 and Hiroyuki Ota 3 We present a basic technology

More information

医系の統計入門第 2 版 サンプルページ この本の定価 判型などは, 以下の URL からご覧いただけます. このサンプルページの内容は, 第 2 版 1 刷発行時のものです.

医系の統計入門第 2 版 サンプルページ この本の定価 判型などは, 以下の URL からご覧いただけます.   このサンプルページの内容は, 第 2 版 1 刷発行時のものです. 医系の統計入門第 2 版 サンプルページ この本の定価 判型などは, 以下の URL からご覧いただけます. http://www.morikita.co.jp/books/mid/009192 このサンプルページの内容は, 第 2 版 1 刷発行時のものです. i 2 t 1. 2. 3 2 3. 6 4. 7 5. n 2 ν 6. 2 7. 2003 ii 2 2013 10 iii 1987

More information

「霧」や「もや」などをクリアにする高速画像処理技術

「霧」や「もや」などをクリアにする高速画像処理技術 Fas Single-Image Defogging 谭志明 白向晖 王炳融 東明浩 あらまし CPU GPU720 48050 fps Absrac Bad weaher condiions such as fog, haze, and dus ofen reduce he performance of oudoor cameras. In order o improve he visibiliy

More information

18 2 20 W/C W/C W/C 4-4-1 0.05 1.0 1000 1. 1 1.1 1 1.2 3 2. 4 2.1 4 (1) 4 (2) 4 2.2 5 (1) 5 (2) 5 2.3 7 3. 8 3.1 8 3.2 ( ) 11 3.3 11 (1) 12 (2) 12 4. 14 4.1 14 4.2 14 (1) 15 (2) 16 (3) 17 4.3 17 5. 19

More information

006 11 8 0 3 1 5 1.1..................... 5 1......................... 6 1.3.................... 6 1.4.................. 8 1.5................... 8 1.6................... 10 1.6.1......................

More information

NLMIXED プロシジャを用いた生存時間解析 伊藤要二アストラゼネカ株式会社臨床統計 プログラミング グループグルプ Survival analysis using PROC NLMIXED Yohji Itoh Clinical Statistics & Programming Group, A

NLMIXED プロシジャを用いた生存時間解析 伊藤要二アストラゼネカ株式会社臨床統計 プログラミング グループグルプ Survival analysis using PROC NLMIXED Yohji Itoh Clinical Statistics & Programming Group, A NLMIXED プロシジャを用いた生存時間解析 伊藤要二アストラゼネカ株式会社臨床統計 プログラミング グループグルプ Survival analysis using PROC NLMIXED Yohji Itoh Clinical Statistics & Programming Group, AstraZeneca KK 要旨 : NLMIXEDプロシジャの最尤推定の機能を用いて 指数分布 Weibull

More information

Vol1-CVIM-172 No.7 21/5/ Shan 1) 2 2)3) Yuan 4) Ancuti 5) Agrawal 6) 2.4 Ben-Ezra 7)8) Raskar 9) Image domain Blur image l PSF b / = F(

Vol1-CVIM-172 No.7 21/5/ Shan 1) 2 2)3) Yuan 4) Ancuti 5) Agrawal 6) 2.4 Ben-Ezra 7)8) Raskar 9) Image domain Blur image l PSF b / = F( Vol1-CVIM-172 No.7 21/5/27 1 Proposal on Ringing Detector for Image Restoration Chika Inoshita, Yasuhiro Mukaigawa and Yasushi Yagi 1 A lot of methods have been proposed for restoring blurred images due

More information

JIS Z803: (substitution method) 3 LCR LCR GPIB

JIS Z803: (substitution method) 3 LCR LCR GPIB LCR NMIJ 003 Agilent 8A 500 ppm JIS Z803:000 50 (substitution method) 3 LCR LCR GPIB Taylor 5 LCR LCR meter (Agilent 8A: Basic accuracy 500 ppm) V D z o I V DUT Z 3 V 3 I A Z V = I V = 0 3 6 V, A LCR meter

More information

(a) (b) (c) Canny (d) 1 ( x α, y α ) 3 (x α, y α ) (a) A 2 + B 2 + C 2 + D 2 + E 2 + F 2 = 1 (3) u ξ α u (A, B, C, D, E, F ) (4) ξ α (x 2 α, 2x α y α,

(a) (b) (c) Canny (d) 1 ( x α, y α ) 3 (x α, y α ) (a) A 2 + B 2 + C 2 + D 2 + E 2 + F 2 = 1 (3) u ξ α u (A, B, C, D, E, F ) (4) ξ α (x 2 α, 2x α y α, [II] Optimization Computation for 3-D Understanding of Images [II]: Ellipse Fitting 1. (1) 2. (2) (edge detection) (edge) (zero-crossing) Canny (Canny operator) (3) 1(a) [I] [II] [III] [IV ] E-mail sugaya@iim.ics.tut.ac.jp

More information

1 Kinect for Windows M = [X Y Z] T M = [X Y Z ] T f (u,v) w 3.2 [11] [7] u = f X +u Z 0 δ u (X,Y,Z ) (5) v = f Y Z +v 0 δ v (X,Y,Z ) (6) w = Z +

1 Kinect for Windows M = [X Y Z] T M = [X Y Z ] T f (u,v) w 3.2 [11] [7] u = f X +u Z 0 δ u (X,Y,Z ) (5) v = f Y Z +v 0 δ v (X,Y,Z ) (6) w = Z + 3 3D 1,a) 1 1 Kinect (X, Y) 3D 3D 1. 2010 Microsoft Kinect for Windows SDK( (Kinect) SDK ) 3D [1], [2] [3] [4] [5] [10] 30fps [10] 3 Kinect 3 Kinect Kinect for Windows SDK 3 Microsoft 3 Kinect for Windows

More information

proc_miru2005_3.dvi

proc_miru2005_3.dvi (MIRU2005) 2005 7 2 2D-3D 812 8581 6 10 1 815 8540 4 9 1 565 0871 2 2 812 8582 3 1 1 E-mal: yum@rvs.s.kyushu-u.ac.jp 2 2 3 1 3 2 2 Fast Marchng Method 2 3 Fast Marchng Method M 2D-3D Regstraton usng 2D

More information

繰り返しゲームの新展開:

繰り返しゲームの新展開: CIRJE-J-65 001 10 001 10 5 New Progress n Repeated Games: Implct Colluson wth Prvate Montorng Htosh Matsushma Faculty of Economcs, Unversty of Tokyo October 5, 001 Abstract The present paper provdes a

More information

. VOCA (Voce Output Communcaton Ads) [], [] [] [] GloveTalkII [] F/F [] [] [8] [9], [0] HMM ) ).. m a m n b n. Stylanou [] a bz=[a, b] (Gaussan Mxture

. VOCA (Voce Output Communcaton Ads) [], [] [] [] GloveTalkII [] F/F [] [] [8] [9], [0] HMM ) ).. m a m n b n. Stylanou [] a bz=[a, b] (Gaussan Mxture THE INSTITUTE OF ELECTRONICS, INFORMATION AND COMMUNICATION ENGINEERS TECHNICAL REPORT OF IEICE. 8 00 E-mal: {kunkosh,qao,mne,hrose}@gavo.t.u-tokyo.ac.jp Nasal sound generaton and ptch control for the

More information

5 Basis Expansions and Regularization 5.3 Filtering and Feature Extraction 5.4 Smoothing Splines Degrees of Freedom and Smoother Matrices

5 Basis Expansions and Regularization  5.3 Filtering and Feature Extraction 5.4 Smoothing Splines Degrees of Freedom and Smoother Matrices Learnng to Generate wth Memory Chongxuan L, Jun Zhu, Bo Zhang 2016/07/16 ICML 読み会 奈良先端科学技術大学院大学 中村研 D2 品川政太朗 1/24 はじめに 何の論文? Neural Turng Machne や Memory Networks で用いられている外部メモリを深層生成モデル (DGMs) に導入 どこがすごいか抽象度の低い特徴は階層が上がるにつれて落ちる情報だが

More information

再発見を試みるユーザ 入力閲覧ページ出力同位ページ 以前に閲覧したページ 同位ページの推定 2. 1 [4], [13] Dubroy [4] [13] 4 [1], [2], [8], [10], [12] Nshmoto [8] Capra [2] Exact Path Su

再発見を試みるユーザ 入力閲覧ページ出力同位ページ 以前に閲覧したページ 同位ページの推定 2. 1 [4], [13] Dubroy [4] [13] 4 [1], [2], [8], [10], [12] Nshmoto [8] Capra [2] Exact Path Su DEIM Forum 2015 B2-5 606 8501 E-mal: {takeda,ohshma,tanaka}@dl.kus.kyoto-u.ac.jp Web 1. Web Web 44% [9] 33% [11] Web 2 3 4 5 6 再発見を試みるユーザ 入力閲覧ページ出力同位ページ 以前に閲覧したページ 同位ページの推定 2. 1 [4], [13] Dubroy [4] [13]

More information

.. F x) = x ft)dt ), fx) : PDF : probbility density function) F x) : CDF : cumultive distribution function F x) x.2 ) T = µ p), T : ) p : x p p = F x

.. F x) = x ft)dt ), fx) : PDF : probbility density function) F x) : CDF : cumultive distribution function F x) x.2 ) T = µ p), T : ) p : x p p = F x 203 7......................................2................................................3.....................................4 L.................................... 2.5.................................

More information

IS1-09 第 回画像センシングシンポジウム, 横浜,14 年 6 月 2 Hough Forest Hough Forest[6] Random Forest( [5]) Random Forest Hough Forest Hough Forest 2.1 Hough Forest 1 2.2

IS1-09 第 回画像センシングシンポジウム, 横浜,14 年 6 月 2 Hough Forest Hough Forest[6] Random Forest( [5]) Random Forest Hough Forest Hough Forest 2.1 Hough Forest 1 2.2 IS1-09 第 回画像センシングシンポジウム, 横浜,14 年 6 月 MI-Hough Forest () E-mail: ym@vision.cs.chubu.ac.jphf@cs.chubu.ac.jp Abstract Hough Forest Random Forest MI-Hough Forest Multiple Instance Learning Bag Hough Forest

More information

(3.6 ) (4.6 ) 2. [3], [6], [12] [7] [2], [5], [11] [14] [9] [8] [10] (1) Voodoo 3 : 3 Voodoo[1] 3 ( 3D ) (2) : Voodoo 3D (3) : 3D (Welc

(3.6 ) (4.6 ) 2. [3], [6], [12] [7] [2], [5], [11] [14] [9] [8] [10] (1) Voodoo 3 : 3 Voodoo[1] 3 ( 3D ) (2) : Voodoo 3D (3) : 3D (Welc 1,a) 1,b) Obstacle Detection from Monocular On-Vehicle Camera in units of Delaunay Triangles Abstract: An algorithm to detect obstacles by using a monocular on-vehicle video camera is developed. Since

More information

untitled

untitled 主成分分析 (Prncpal Component Analy) で情報を集約する マルチスペクトル画像 なし が情報を集約する 69.68 77.97 85.73 96.7 98.8 画像 : NASA 除去できる一部に集約 あり.24.35 4.63 7.65 3.9 分散の比率 最大を 255, 最小を に正規化して表示 3 つの成分から画像を再生した 信号処理の手順 行列 A 共分散行列に対する

More information

基底関数ネットワーク

基底関数ネットワーク 6. 基底関数ネットワーク (Bass Functon Network) 6-1 基底関数ネットワーク研究の背景 (1)( 階層型 ) ニューラルネットワークの問題点の回避 設計性の悪さ ローカルミニマム問題 (2) 級数展開の利用 基底関数が周期関数 フーリエ級数 フーリエ級数 フーリエ級数 F1 フーリエ係数 F2 信号 + F3 F4 フーリエ展開で関数を近似した例 フーリエ係数の意味 F1

More information

IPSJ SIG Technical Report Vol.2012-MUS-96 No /8/10 MIDI Modeling Performance Indeterminacies for Polyphonic Midi Score Following and

IPSJ SIG Technical Report Vol.2012-MUS-96 No /8/10 MIDI Modeling Performance Indeterminacies for Polyphonic Midi Score Following and MIDI 1 2 3 2 1 Modeling Performance Indeterminacies for Polyphonic Midi Score Following and Its Application to Automatic Accompaniment Nakamura Eita 1 Yamamoto Ryuichi 2 Saito Yasuyuki 3 Sako Shinji 2

More information

1 (1) 3 4 (2) 5 () 30 35 40 45 50 55 60 410 519 599 975 1,176 1,362 1,840 250 298 426 691 735 829 693 865 1,287 1,838 2,209 2,718 2,776 3,361 888 1,061 1,448 1,598 1,849 2,157 2,214 1,324 1,172 964 808

More information

LLG-R8.Nisus.pdf

LLG-R8.Nisus.pdf d M d t = γ M H + α M d M d t M γ [ 1/ ( Oe sec) ] α γ γ = gµ B h g g µ B h / π γ g = γ = 1.76 10 [ 7 1/ ( Oe sec) ] α α = λ γ λ λ λ α γ α α H α = γ H ω ω H α α H K K H K / M 1 1 > 0 α 1 M > 0 γ α γ =

More information

untitled

untitled 3 3. (stochastic differential equations) { dx(t) =f(t, X)dt + G(t, X)dW (t), t [,T], (3.) X( )=X X(t) : [,T] R d (d ) f(t, X) : [,T] R d R d (drift term) G(t, X) : [,T] R d R d m (diffusion term) W (t)

More information

タイトル

タイトル Flud Flow Smulton wth Cellulr Automt 00N2100008J 2002 225 1. cellulr utomton n t+ 1 t t = f( r, L, + r (1 t t+ 1 f t r t +1 Prllel Vrtul Mchne Messge-Pssng Interfce 1 2. 2. 1 t t 0 t = 1 = 1 = t = 2 2

More information

TERG

TERG Dscusson Paper No. 268 小標本特性に優れたパネル単位根検定 千木良弘朗 山本拓 2011 年 7 月 TOHOKU ECONOMICS RESEARCH GROUP GRADUATE SCHOOL OF ECONOMICS AND MANAGEMENT TOHOKU UNIVERSITY KAWAUCHI, AOBA-KU, SENDAI, 980-8576 JAPAN Λ z

More information

08_研究速報_3校_1002.indd

08_研究速報_3校_1002.indd 196 MEDICAL IMAGING TECHNOLOGY Vol. 36 No. 4 September 2018 MRI 1 2 2 1 3 4 1 MRI MRI ScSR MRI T1 T2 FLAIR TOF ScSR PSNR SSIM T1 T1 T2 FLAIR ScSR PSNR SSIM Bilinear Bicubic Lanczos p

More information

EQUIVALENT TRANSFORMATION TECHNIQUE FOR ISLANDING DETECTION METHODS OF SYNCHRONOUS GENERATOR -REACTIVE POWER PERTURBATION METHODS USING AVR OR SVC- Ju

EQUIVALENT TRANSFORMATION TECHNIQUE FOR ISLANDING DETECTION METHODS OF SYNCHRONOUS GENERATOR -REACTIVE POWER PERTURBATION METHODS USING AVR OR SVC- Ju EQUIVALENT TRANSFORMATION TECHNIQUE FOR ISLANDING DETECTION METHODS OF SYNCHRONOUS GENERATOR -REACTIVE POWER PERTURBATION METHODS USING AVR OR SVC- Jun Motohashi, Member, Takashi Ichinose, Member (Tokyo

More information

IPSJ SIG Technical Report Vol.2009-DBS-149 No /11/ Bow-tie SCC Inter Keyword Navigation based on Degree-constrained Co-Occurrence Graph

IPSJ SIG Technical Report Vol.2009-DBS-149 No /11/ Bow-tie SCC Inter Keyword Navigation based on Degree-constrained Co-Occurrence Graph 1 2 1 Bow-tie SCC Inter Keyword Navigation based on Degree-constrained Co-Occurrence Graph Satoshi Shimada, 1 Tomohiro Fukuhara 2 and Tetsuji Satoh 1 We had proposed a navigation method that generates

More information

2. CABAC CABAC CABAC 1 1 CABAC Figure 1 Overview of CABAC 2 DCT 2 0/ /1 CABAC [3] 3. 2 値化部 コンテキスト計算部 2 値算術符号化部 CABAC CABAC

2. CABAC CABAC CABAC 1 1 CABAC Figure 1 Overview of CABAC 2 DCT 2 0/ /1 CABAC [3] 3. 2 値化部 コンテキスト計算部 2 値算術符号化部 CABAC CABAC H.264 CABAC 1 1 1 1 1 2, CABAC(Context-based Adaptive Binary Arithmetic Coding) H.264, CABAC, A Parallelization Technology of H.264 CABAC For Real Time Encoder of Moving Picture YUSUKE YATABE 1 HIRONORI

More information

インターリーブADCでのタイミングスキュー影響のデジタル補正技術

インターリーブADCでのタイミングスキュー影響のデジタル補正技術 1 インターリーブADCでのタイミングスキュー影響のデジタル補正技術 浅見幸司 黒沢烈士 立岩武徳 宮島広行 小林春夫 ( 株 ) アドバンテスト 群馬大学 2 目次 1. 研究背景 目的 2. インターリーブADCの原理 3. チャネル間ミスマッチの影響 3.1. オフセットミスマッチの影響 3.2. ゲインミスマッチの影響 3.3. タイミングスキューの影響 4. 提案手法 4.1. インターリーブタイミングミスマッチ補正フィルタ

More information

IPSJ SIG Technical Report Vol.2011-EC-19 No /3/ ,.,., Peg-Scope Viewer,,.,,,,. Utilization of Watching Logs for Support of Multi-

IPSJ SIG Technical Report Vol.2011-EC-19 No /3/ ,.,., Peg-Scope Viewer,,.,,,,. Utilization of Watching Logs for Support of Multi- 1 3 5 4 1 2 1,.,., Peg-Scope Viewer,,.,,,,. Utilization of Watching Logs for Support of Multi-View Video Contents Kosuke Niwa, 1 Shogo Tokai, 3 Tetsuya Kawamoto, 5 Toshiaki Fujii, 4 Marutani Takafumi,

More information

ベイズ統計入門

ベイズ統計入門 ベイズ統計入門 条件付確率 事象 F が起こったことが既知であるという条件の下で E が起こる確率を条件付確率 (codtoal probablt) という P ( E F ) P ( E F ) P( F ) 定義式を変形すると 確率の乗法公式となる ( E F ) P( F ) P( E F ) P( E) P( F E) P 事象の独立 ある事象の生起する確率が 他のある事象が生起するかどうかによって変化しないとき

More information