• Search Research Projects
  • Search Researchers
  • How to Use
  1. Back to previous page

Discriminant analysis of spatial data based on fusion of Markov random fields and machine learning

Research Project

Project/Area Number 15540123
Research Category

Grant-in-Aid for Scientific Research (C)

Allocation TypeSingle-year Grants
Section一般
Research Field General mathematics (including Probability theory/Statistical mathematics)
Research InstitutionKYUSHU UNIVERSITY (2004)
Hiroshima University (2003)

Principal Investigator

NISHII Ryuei  Kyushu University, Faculty of Mathematics, Professor, 大学院・数理学研究院, 教授 (40127684)

Co-Investigator(Kenkyū-buntansha) ASANO Akira  Hiroshima University, Faculty of Integrated Arts and Sciences, Associate Professor, 総合科学部, 助教授 (60243987)
IIKURA Yoshikazu  Hirosaki University, Faculty of Science and Engineering, Professor, 理工学部, 教授 (30109897)
EGUCHI Shinto  Kyushu University, Institute of Statistical Mathematics, Professor, 統計基礎研究系, 教授 (10168776)
SAKATA Toshio  Kyushu University, Faculty of Design, Professor, 大学院・芸術工学研究院, 教授 (20117352)
TANAKA Shojiro  Shimane University, Interdisciplinary, Faculty of Science and Engineering, Professor, 総合理工学部, 教授 (00197427)
高野 秀敏  広島大学, 総合科学部, 教務員 (20281157)
Project Period (FY) 2003 – 2004
Project Status Completed (Fiscal Year 2004)
Budget Amount *help
¥3,600,000 (Direct Cost: ¥3,600,000)
Fiscal Year 2004: ¥1,700,000 (Direct Cost: ¥1,700,000)
Fiscal Year 2003: ¥1,900,000 (Direct Cost: ¥1,900,000)
KeywordsAdaBoost / Discriminant analysis / Exponential loss function / Image segmentation / Machine learning / Markov random fields / Exponentiak loss function / 学習理論 / 判別分析 / SVM / アダブースト / 多重分光画像 / マイクロアレイデータ / 文字認識
Research Abstract

Aims of the research were as follows : (a)Improvement of multispectral image classification due to machine learning based on spatial distributions of categories, (b)Derivation of selection method of machine learning approaches and model selection of Markov random fields, (c)Comparison of the proposed and the ordinary methods through actual satellite images. Hence the aims were to derive a fused classification method based on statistics and machine learning in short. Nishii and Eguchi (2004) proposed Spatial AdaBoost, which achieved the aims.
Spatial AdaBoost is carried out in the following steps, (1)Obtain the posterior probabilities of the training data. (2)Calculate averages of log posteriors in neighborhoods of each pixels, and regard them as classification functions. (3)Tune the weights for the log posteriors by minimizing the exponential risk sequentially. (4)Classify test data by the convex combination of log posteriors.
The proposed method is examined through simulated and real data sets, and it is seen that the method is very fast and shows a similar performance to MRF-based classifier. The method sometimes gives a negative weight for the log posterior for some case because the exponential loss puts a huge penalty for outlying misclassified data. Hence, Spatial AdaBoost based on robust loss functions is under investigation, and we obtain a partial answer.
Further, we studied issues related to image analysis and machine learning, for example, robust loss functions and properties ; corrections of geometric and topographic effects ; morphological texture analysis ; and derivation of rotation invariant moments for character recognition.

Report

(3 results)
  • 2004 Annual Research Report   Final Research Report Summary
  • 2003 Annual Research Report
  • Research Products

    (33 results)

All 2004 2003 Other

All Journal Article (27 results) Publications (6 results)

  • [Journal Article] A new series of rotation invariant moments derived by Lie transformation group theory2004

    • Author(s)
      T.Sakata
    • Journal Title

      Proc.of 9th International Workshop on Frontiers in Handwriting Recognition WFHR-9

      Pages: 377-381

    • Description
      「研究成果報告書概要(和文)」より
    • Related Report
      2004 Final Research Report Summary
  • [Journal Article] A paradox concerning nuisance parameters and projected estimating functions2004

    • Author(s)
      M.Henmi
    • Journal Title

      Biometrika 91

      Pages: 929-941

    • Description
      「研究成果報告書概要(和文)」より
    • Related Report
      2004 Final Research Report Summary
  • [Journal Article] Robust principal component analysis with adaptive selection for tuning parameters2004

    • Author(s)
      I.Higuchi
    • Journal Title

      J.Machine Learning Research 5

      Pages: 453-471

    • Description
      「研究成果報告書概要(和文)」より
    • Related Report
      2004 Final Research Report Summary
  • [Journal Article] Supervised image classification based on AdaBoost with contextual weak classifiers2004

    • Author(s)
      R.Nishii
    • Journal Title

      Proc.Internat.Geoscience & Remote Sensing Sympo.2004 II

      Pages: 1467-1470

    • Description
      「研究成果報告書概要(和文)」より
    • Related Report
      2004 Final Research Report Summary
  • [Journal Article] Robustifying AdaBoost by adding the naive error rate2004

    • Author(s)
      T.Takenouchi
    • Journal Title

      Neural Computation 16

      Pages: 767-787

    • Description
      「研究成果報告書概要(和文)」より
    • Related Report
      2004 Final Research Report Summary
  • [Journal Article] 海と陸の標本点の簡便な識別に基づく幾何補正の最適化2004

    • Author(s)
      飯倉 善和
    • Journal Title

      日本リモートセンシング学会誌 24

      Pages: 378-386

    • NAID

      10013792900

    • Description
      「研究成果報告書概要(和文)」より
    • Related Report
      2004 Final Research Report Summary
  • [Journal Article] A new series of rotation invariant moments derived by Lie transformation group theory2004

    • Author(s)
      T.Sakata, R.Nishii, T-S.Chin, R.Sawae
    • Journal Title

      Proc.of 9th International Workshop on Frontiers in Handwriting Recognition WFHR-9

      Pages: 377-381

    • Description
      「研究成果報告書概要(欧文)」より
    • Related Report
      2004 Annual Research Report 2004 Final Research Report Summary
  • [Journal Article] A paradox concerning nuisance parameters and projected estimating functions2004

    • Author(s)
      M.Henmi, T.Takenouchi, T.Kanamori, S.Eguchi
    • Journal Title

      Biometrika 91

      Pages: 929-941

    • Description
      「研究成果報告書概要(欧文)」より
    • Related Report
      2004 Final Research Report Summary
  • [Journal Article] Robust principal component analysis with adaptive selection for tuning parameters2004

    • Author(s)
      I.Higuchi, S.Eguchi
    • Journal Title

      J.Machine Learning Research 5

      Pages: 453-471

    • Description
      「研究成果報告書概要(欧文)」より
    • Related Report
      2004 Annual Research Report 2004 Final Research Report Summary
  • [Journal Article] Supervised image classification based on AdaBoost with contextual weak classifiers2004

    • Author(s)
      R.Nishii, S.Eguchi
    • Journal Title

      Proc.Internat.Geoscience & Remote Sensing Sympo. 2004, II

      Pages: 1467-1470

    • Description
      「研究成果報告書概要(欧文)」より
    • Related Report
      2004 Final Research Report Summary
  • [Journal Article] Robustifying AdaBoost by adding the naive error rate2004

    • Author(s)
      T.Takenouchi, S.Eguchi
    • Journal Title

      Neural Computation 16

      Pages: 767-787

    • Description
      「研究成果報告書概要(欧文)」より
    • Related Report
      2004 Final Research Report Summary
  • [Journal Article] Optimization of geometric correction of satellite images based on a simple discrimination of samples from sea and land2004

    • Author(s)
      Y.Iikura, H.Saito, Y.Tobita
    • Journal Title

      J.Remote Sensing Soc.Japan 24

      Pages: 378-386

    • NAID

      10013792900

    • Description
      「研究成果報告書概要(欧文)」より
    • Related Report
      2004 Final Research Report Summary
  • [Journal Article] Supervised image classification based on AdaBoost with contextual weak classifiers2004

    • Author(s)
      R.Nishii, S.Eguchi
    • Journal Title

      Proc.of 2004 IEEE International Geoscience and Remote Sensing Symposium II

      Pages: 1467-1470

    • Related Report
      2004 Annual Research Report
  • [Journal Article] Deforestation models due to population and relief energy in east Asia2004

    • Author(s)
      S.Tanaka, R.Nishii
    • Journal Title

      Proc.of 8th China-Japan symposium on statistics

      Pages: 272-275

    • Related Report
      2004 Annual Research Report
  • [Journal Article] A paradox concerning nuisance parameters and projected estimating functions2004

    • Author(s)
      M.Henmi, S.Eguchi
    • Journal Title

      Biometrika 91

      Pages: 929-941

    • Related Report
      2004 Annual Research Report
  • [Journal Article] 白神山地の環境解析のための衛星画像処理2004

    • Author(s)
      飯倉善和
    • Journal Title

      白神研究 創刊号

      Pages: 14-23

    • NAID

      110002555894

    • Related Report
      2004 Annual Research Report
  • [Journal Article] A Markov random field-based approach to decision level fusion for remote sensing image classification2003

    • Author(s)
      R.Nishii
    • Journal Title

      IEEE Transactions on Geoscience and Remote Sensing 41

      Pages: 2316-2319

    • Description
      「研究成果報告書概要(和文)」より
    • Related Report
      2004 Final Research Report Summary
  • [Journal Article] 人口増加に伴う森林減少の空間モデル2003

    • Author(s)
      田中 章司郎
    • Journal Title

      応用統計学 32

      Pages: 1-15

    • NAID

      10011680014

    • Description
      「研究成果報告書概要(和文)」より
    • Related Report
      2004 Final Research Report Summary
  • [Journal Article] Contextual Image Segmentation} based on AdaBoost and Markov Random Fields2003

    • Author(s)
      R.Nishii
    • Journal Title

      Proc.Internat.Geoscience & Remote Sensing Sympo.2003 VI

      Pages: 3507-3509

    • Description
      「研究成果報告書概要(和文)」より
    • Related Report
      2004 Final Research Report Summary
  • [Journal Article] Morphological texture analysis using optimization of structuring elements2003

    • Author(s)
      A.Asano
    • Journal Title

      Geometry, Morphology, and Computational Imaging, Springer LNCS 2616

      Pages: 141-152

    • Description
      「研究成果報告書概要(和文)」より
    • Related Report
      2004 Final Research Report Summary
  • [Journal Article] Local likelihood method and theory for a bridge between parametric and nonparametric regression.2003

    • Author(s)
      S.Eguchi
    • Journal Title

      J.Nonparametric Statistics 15

      Pages: 665-683

    • Description
      「研究成果報告書概要(和文)」より
    • Related Report
      2004 Final Research Report Summary
  • [Journal Article] 放射照度の空間的な変動を考慮した地形効果補正2003

    • Author(s)
      飯倉 善和
    • Journal Title

      日本リモートセンシング学会誌 23

      Pages: 386-392

    • Description
      「研究成果報告書概要(和文)」より
    • Related Report
      2004 Final Research Report Summary
  • [Journal Article] Models of deforestation with spatial dependency by human population interactions2003

    • Author(s)
      S.Tanaka, R.Nishii
    • Journal Title

      Japanese J.Appl.Statist. 32

      Pages: 1-15

    • NAID

      10011680014

    • Description
      「研究成果報告書概要(欧文)」より
    • Related Report
      2004 Final Research Report Summary
  • [Journal Article] Contextual image segmentation based on AdaBoost and Markov random fields2003

    • Author(s)
      R.Nishii
    • Journal Title

      Proc.Internat.Geoscience & Remote Sensing Sympo. 2003, VI

      Pages: 3507-3509

    • Description
      「研究成果報告書概要(欧文)」より
    • Related Report
      2004 Final Research Report Summary
  • [Journal Article] Morphological texture analysis using optimization of structuring elements2003

    • Author(s)
      A.Asano, M.Miyagawa, M.Fujio
    • Journal Title

      Geometry, Morphology, and Computational Imaging(Springer LNCS) 2616

      Pages: 141-152

    • Description
      「研究成果報告書概要(欧文)」より
    • Related Report
      2004 Final Research Report Summary
  • [Journal Article] Local likelihood method and theory for a bridge between parametric and nonparametric regression.2003

    • Author(s)
      S.Eguchi, T.Y.Kim, B.U.Park
    • Journal Title

      J.Nonparametric Statistics 15

      Pages: 665-683

    • Description
      「研究成果報告書概要(欧文)」より
    • Related Report
      2004 Final Research Report Summary
  • [Journal Article] Correction of topographic effects in satellite images with due consideration on spatial variation of irradiance2003

    • Author(s)
      Y.Iikura, H.Saito, S.Tanba
    • Journal Title

      J.Remote Sensing Soc.Japan 23

      Pages: 386-392

    • NAID

      10011818434

    • Description
      「研究成果報告書概要(欧文)」より
    • Related Report
      2004 Final Research Report Summary
  • [Publications] R.Nishii: "A Markov random field-based approach to decision level fusion for remote sensing image classification"IEEE Transactions on Geoscience and Remote Sensing. 41・10. 2316-2319 (2003)

    • Related Report
      2003 Annual Research Report
  • [Publications] 田中 章司郎: "人口増加に伴う森林減少の空間モデル"応用統計学. 32・1. 1-15 (2003)

    • Related Report
      2003 Annual Research Report
  • [Publications] R.Nishii: "Contextual Image Segmentation} based on AdaBoost and Markov Random Fields"Proc.Internal.Geoscience & Remote Sensing Sympo.2003. VI. 3507-3509 (2003)

    • Related Report
      2003 Annual Research Report
  • [Publications] T.Takenouchi: "Robustifying AdaBoost by adding the naive error rate"Neural Computation. 16(To appear). (2004)

    • Related Report
      2003 Annual Research Report
  • [Publications] A.Asano: "Morphological texture analysis using optimization of structuring elements"Geometry, Morphology, and Computational Imaging, Springer LNCS. 2616. 141-152 (2003)

    • Related Report
      2003 Annual Research Report
  • [Publications] S.Eguchi: "Local likelihood method and theory for a bridge between parametric and nonparametric regression"J.Nonparametric Statistics. 15. 665-683 (2003)

    • Related Report
      2003 Annual Research Report

URL: 

Published: 2003-04-01   Modified: 2016-04-21  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi