Testes Estatísticos

Este seção compara as melhores versões dos descritores, considerando a hora do dia, o canal de cor e o threshold usado para binarizar à imagem de recurrence plot. São elas: Edge Histogram (thre. 10, 6h, Canal B), HOG (thre. 20, 6h, Canal B), Moments (thre. 5, 6h, Canal B) e PHOG (thre. 10, 6h, Canal B).

Tabela 4.1: Diferenças de P @5 entre as melhores versões de descritores de forma, consi- derando hora do dia, canal de cor e threshold.

Intervalo de Confiança (95%)

Método min. max.

Edge Histogram (thre. 10, 6h, Canal B) – HOG (thre. 20, 6h, Canal B) -0.146 0.006

Edge Histogram (thre. 10, 6h, Canal B) – Moments (thre. 5, 6h, Canal B) 0,064 0,215

Edge Histogram (thre. 10, 6h, Canal B) – PHOG (thre. 10, 6h, Canal B) -0,195 0,015

HOG (thre. 20, 6h, Canal B) – Moments (thre. 5, 6h, Canal B) 0,085 0,117

HOG (thre. 20, 6h, Canal B) – PHOG (thre. 10, 6h, Canal B) -0,003 0,028

Moments (thre. 5, 6h, Canal B) – PHOG (thre. 10, 6h, Canal B) -0,103 -0,0734

A Tabela 4.1 apresenta os resultados do intervalo de confiança da diferença entre os melhores métodos comparados. Como pode ser observado o Edge Histogram tem resultado médio da ordem de 0, 38 e o HOG de 0, 32. Mas não há diferença estatística entre eles. O mesmo é observado em relação aos resultados do PHOG quando comparados aos do Edge Histogram e o HOG. Pode ser observado também que todos eles apresentam resultados estatisticamente superiores ao do Moments Invariants.

Capítulo 5

Conclusões

Esta dissertação tratou da representação de séries temporais extraídas a partir dos dados fenológicos do projeto e-phenology usando a representação recurrence plot (RP). O RP é uma representação capaz de revelar em quais pontos algumas trajetórias retornam a um estado visitado anteriormente. O objetivo deste trabalho identificar e investigar descrito- res de forma adequados para a caracterização de séries temporais associadas a diferentes horas do dia, a diferentes canais de cor, e a diferentes áreas de interesse. Trata-se de primeiro trabalho voltado à caracterização de imagens RP usando descritores de forma.

Os resultados experimentais apontam para a boa eficácia de descritores de forma na caracterização de representações RP de séries temporais, o que mostra que seu uso é promissor no contexto de busca de indivíduos de mesma espécie (regiões de interesse) em imagens de vegetação. Com base em nossos experimentos, constatamos que:

• em geral, o HOG apresentou os melhores resultados quando comparado aos outros descritores na grande maioria das horas do dia e nos três canais de cor.

• o descritor Moments Invariants em geral teve o pior resultado para todos os canais de cor e todas as horas do dia;

• em geral, os descritores tiveram o melhor desempenho próximo as horas extremas do dia, ou seja, próximo às 06:00 e próximo às 18:00h;

• os descritores apresentaram melhores desempenho com os valores de limiar 20 e 50 na Equação 2.1.

• no geral, os descritores apresentaram os melhores resultados no canal B. Como trabalhos futuros pretendemos:

1. investigar outros descritores de forma recentemente propostos na caracterização de imagens recurrence plot [61, 22];

2. estender a avaliação realizada para séries temporais de outros domínios (séries do mercado financeiro, por exemplo) e ou séries fenológicas associadas a dados de campo [3].

CAPÍTULO 5. CONCLUSÕES 45

3. investigar a correlação dos descritores de forma visando a definir mecanismos apro- priados para combiná-los. Técnicas supervisionadas [36, 18, 17] e não supervisiona- das [41] poderiam ser empregadas neste processo. Uma outra vertente associada a este tópico consiste na combinação de descritores de forma com outros tipos de des- critores (por exemplo, descritores de textura). Um trabalho promissor para iniciar esta linha de pesquisa consiste no survey apresentado em [42].

Referências Bibliográficas

[1] H. Ahrends, S. Etzold, W. Kutsch, R. Stoeckli, R. Bruegger, F. Jeanneret, H. Wanner, N. Buchmann, and W. Eugster. Tree phenology and carbon dioxide fluxes: use of digital photography at for process-based interpretation the ecosystem scale. Climate Research, 39:261–274, 2009.

[2] H. Ellen Ahrends, R. Brügger, Reto Stöckli, Jürg Schenk, Pavel Michna, Francois Je- anneret, Heinz Wanner, and Werner Eugster. Quantitative phenological observations of a mixed beech forest in northern switzerland with digital photography. Journal of Geophysical Research: Biogeosciences, 113(G4):n/a–n/a, 2008. G04004.

[3] B. Alberton, J. Almeida, R. Helm, R. S. Torres, A. Menzel, and L. P. C. Morellato. Using phenological cameras to track the green up in a cerrado savanna and its on- the-ground validation. Ecological Informatics, 19:62 – 70, 2014.

[4] J. Almeida, J. A. Santos, B. Alberton, L. P. C. Morellato, and R. S. Torres. Vi- sual rhythm-based time series analysis for phenology studies. In IEEE International Conference on Image Processing, ICIP 2013, Melbourne, Australia, September 15-18, 2013, pages 4412–4416, 2013.

[5] J. Almeida, J. A. Santos, B. Alberton, R. S. Torres, and L. P. C. Morellato. Remote phenology: Applying machine learning to detect phenological patterns in a cerrado savanna. In IEEE International Conference on eScience (eScience’12), pages 1–8, 2012.

[6] J. Almeida, J. A. Santos, B. Alberton, R. S. Torres, and L. P. C. Morellato. Applying machine learning based on multiscale classifiers to detect remote phenology patterns in cerrado savanna trees. Ecological Informatics, 23:49 – 61, 2014. Special Issue on Multimedia in Ecology and Environment.

[7] F. A. Andaló. Descritores de forma baseados em tensor scale. Msc thesis, Universi- dade Estadual de Campinas . Instituto de Computação., 2007.

[8] A. Bosch, A. Zisserman, and X. Munoz. Representing shape with a spatial pyramid kernel. In Proceedings of the 6th ACM International Conference on Image and Video Retrieval, CIVR ’07, pages 401–408, New York, NY, USA, 2007. ACM.

[9] S. Chang, T. Sikora, and A. Purl. Overview of the mpeg-7 standard. Circuits and Systems for Video Technology, IEEE Transactions on, 11(6):688–695, Jun 2001.

REFERÊNCIAS BIBLIOGRÁFICAS 47

[10] C. Chatfield. The analysis of time series: an introduction. CRC Press, Florida, US, 6th edition, 2004.

[11] M. A. Crimmins and T. M. Crimmins. Monitoring plant phenology using digital repeat photography. Environmental Management, 41(6):949–958, 2008.

[12] N. Dalal and B. Triggs. Histograms of oriented gradients for human detection. In Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’05) - Volume 1 - Volume 01, CVPR ’05, pages 886–893, Washington, DC, USA, 2005. IEEE Computer Society.

[13] S. A. Dudani, K. J. Breeding, and R. B. McGhee. Aircraft identification by moment invariants. Computers, IEEE Transactions on, C-26(1):39–46, Jan 1977.

[14] J. P. Eckmann, O. S. Kamphorst, and D. Ruelle. Recurrence plots of dynamical systems. Europhysics Letters, vol. 4, no. 9:973–977, 1987.

[15] J.-P. Eckmann, S. Oliffson Kamphorst, and D. Ruelle. Recurrence plots of dynamical systems. EPL (Europhysics Letters), 4(9):973, 1987.

[16] M. Eom and Y. Choe. Fast extraction of edge histogram in dct domain based on mpeg7. In Proceedings of world academy of science, engineering and technology, volume 9, pages 209–212. Citeseer, 2007.

[17] F. A. Faria, J. Almeida, B. Alberton, L. P. C. Morellato, A. Rocha, and R. S. Torres. Time series-based classifier fusion for fine-grained plant species recognition. Pattern Recognition Letters, pages –, 2015.

[18] F. A. Faria, J. Almeida, B. Alberton, L. P. C. Morellato, and R. S. Torres. Fusion of time series representations for plant recognition in phenology studies. Pattern Recognition Letters, pages –, 2016.

[19] E. A. Graham, E. C. Riordan, M. Yuen, E, D. Estrin, and P. W. Rundel. Public internet-connected cameras used as a cross-continental ground-based plant phenology monitoring system. Global Change Biology, 16(11):3014–3023, 2010.

[20] E. A. Graham, E. M. Yuen, G. F. Robertson, W. J. Kaiser, M. P. Hamilton, and P. W. Rundel. Budburst and leaf area expansion measured with a novel mobile camera system and simple color thresholding. Environmental and Experimental Botany, 65(2–3):238 – 244, 2009.

[21] L. Guigues, J. Cocquerez, and H. Le Men. Scale-sets image analysis. International Journal of Computer Vision, 68:289–317, 2006.

[22] R. A. Guler, S. Tari, and G. Unal. Landmarks inside the shape: Shape matching using image descriptors. Pattern Recognition, 49:79 – 88, 2016.

[23] M. K. Hu. Visual pattern recognition by moment invariants. Information Theory, IRE Transactions on, 8(2):179–187, February 1962.

REFERÊNCIAS BIBLIOGRÁFICAS 48

[24] R. Ide and H. Oguma. Use of digital cameras for phenological observations. Ecological Informatics, 5:339–347, 2010.

[25] J. S. Iwanski and E. Bradley. Recurrence plots of experimental data: To embed or not to embed? Chaos: An Interdisciplinary Journal of Nonlinear Science, 8:861–871, 1998.

[26] V. Kannan and R. Balasubramani. Efficient use of MPEG-7 color layout and edge histogram descriptors in CBIR systems. Global Journal of Computer Science and Technology, 9(4):157–163, 2009.

[27] C. D. Keeling, J. F. S. Chin, and T. P. Whorf. Increased activity of northern vege- tation inferred from atmospheric co2 measurements. Nature, 382:146–149, 1996. [28] L. Keyes and A. Winstanley. Using moment invariants for classifying shapes on large-

scale maps. Computers, Environment and Urban Systems, 25(1):119–130, 2001. [29] S. Kurc and L. Benton. Digital image-derived greenness links deep soil moisture to

carbon uptake in a creosotebush-dominated shrubland. Journal of Arid Environ- ments, 74:585–594, 2010.

[30] S. Lazebnik, C. Schmid, and J. Ponce. Beyond bags of features: Spatial pyramid matching for recognizing natural scene categories. In Proceedings of the 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition - Vo- lume 2, CVPR ’06, pages 2169–2178, Washington, DC, USA, 2006. IEEE Computer Society.

[31] S. Loncaric. A survey of shape analysis techniques. Pattern Recognition, 31(8):983– 1190, 1988.

[32] D. Loustau, A. Bosc, A. Colin, H. Davi, C. François, E. Dufrêne, M. Équé, E. Cloppet, D. Arrouays, C. Le Bas, N. Saby, G. Pignard, N. Hamza, A. Granier, N. Breda, P. Ciais, N. Viovy, J. Ogée, and J. Delage. Modeling the climate change effects on the potential reduction of french plains forests at the sub regional level. Tree Physiol, 25:813–823, 2005.

[33] L. E. Mariote, C. B. Medeiros, R. S. Torres, and L. M. Bueno. TIDES - a new descriptor for time series oscillation behavior. GeoInformatica, 15(1):75–109, 2011. [34] M. Migliavacca, M. Galvagno, E. Cremonese, M. Rossini, M. Meroni, O. Sonnentag,

S. Cogliati, G. Manca, F. Diotri, L. Busetto, . Cescatti, R. Colombo, F. Fava, U. M. Cella, E. Pari, C. Siniscalco, and A. D. Richardson. Using digital repeat photography and eddy covariance data to model grassland phenology and photosynthetic {CO2} uptake. Agricultural and Forest Meteorology, 151(10):1325 – 1337, 2011.

[35] J. T. Morisette, A. D. Richardson, A. K. Knapp, J. I. Fisher, E. A. Graham, J. Abat- zoglou, B. E. Wilson, D. D. Breshears, G. M. Henebry, J. M. Hanes, and L. Liang. Tracking the rhythm of the seasons in the face of global change: phenological research in the 21st century. Frontiers in Ecology and the Environment, 7(5):253–260, 2009.

REFERÊNCIAS BIBLIOGRÁFICAS 49

[36] J. V. Muñoz, R. S. Torres, and M. A. Alves. A soft computing approach for learning to aggregate rankings. In Proceedings of the 24th ACM International on Conference on Information and Knowledge Management, CIKM 2015, Melbourne, VIC, Australia, October 19 - 23, 2015, pages 83–92, 2015.

[37] S. Nagai, T. Maeda, M. Gamo, H. Muraoka, R. Suzuki, and K. N. Nasahara. Using digital camera images to detect canopy condition of deciduous broad-leaved trees. Plant Ecology e Diversity, 4:79–89, 2011.

[38] G. C. S. Negi. Leaf and bud demography and shoot growth in evergreen and deciduous trees of central himalaya. India, 20:416–429, 2006.

[39] C. Parmesan and G. Yohe. A globally coherent fingerprint of climate change impacts across natural systems. Nature, 421:37–42, 2003.

[40] C. Parmesan and G. A. Yohe. A globally coherent fingerprint to climate change impacts accross natural systems. Nature, 421:37–42, 2003.

[41] D. C. G. Pedronette, O. A. B. Penatti, and R. S. Torres. Unsupervised manifold learning using reciprocal knn graphs in image re-ranking and rank aggregation tasks. Image Vision Comput., 32(2):120–130, 2014.

[42] O. A. B. Penatti, E. Valle, and R. S. Torres. Comparative study of global color and texture descriptors for web image retrieval. J. Visual Communication and Image Representation, 23(2):359–380, 2012.

[43] J. Peñuelas and M. Boada. A global change-induced biome shift in the montseny mountains (ne spain). Global Change Biology, 9(2):131–140, 2003.

[44] F. R. Pirolla. Redução de dimensionalidade usando agrupamento e discretização ponderada para a recuperação de imagens por conteúdo. Msc thesis, Universidade de São carlos, Outubro 2012.

[45] R. J. Prokop and A. P. Reeves. A survey of moment-based techniques for unocclu- ded object representation and recognition. CVGIP: Graph. Models Image Process., 54(5):438–460, September 1992.

[46] T. Rakthanmanon and E. Keogh. Fast shapelets: A scalable algorithm for discovering time series shapelets. pages 668–676, 2013.

[47] B. Rathcke and E. P. Lacey. Phenological patterns of terrestrial plants. Annual Review of Ecology and Systematics, 16:179–214, 1985.

[48] E. P. Rathcke, B. Lacey. Phenological patterns of terrestrial plants. Annual Review of Ecology and Systematics,, 16:179–214, 1985.

[49] P. B. Reich. Phenology of tropical forests: patterns, causes and consequences. Ca- nadian Journal of Botany, 73:164–174, 1985.

REFERÊNCIAS BIBLIOGRÁFICAS 50

[50] A. D. Richardson, B. H. Braswell, D. Y. Hollinger, J. P. Jenkins, and S. V. Ollinger. Near-surface remote sensing of spatial and temporal variation in canopy phenology. Ecological Applications, 19:1417–1428, 2009.

[51] A. D. Richardson, J. P. Jenkins, B. H. Braswell, D. Y. Hollinger, S. V. Ollinger, and M. L. Smith. Use of digital webcam images to track spring green-up in a deciduous broadleaf forest. Oecologia, 152:323–334, 2007.

[52] A. D. Richardson, J. P. Jenkins, B. H. Braswell, D. Y. Hollinger, S. V. Ollinger, and M. L. Smith. Use of digital webcam images to track spring greep-up in a deciduous broadleaf forest. Oecologia, 152:323–334, 2007.

[53] C. Rosenzweig, D. Karoly, M. Vicarelli, P. Neofotis, Q. Wu, G. Casassa, A. Menzel, T. L. Root, N. Estrella, B. Seguin, P. Tryjanowski, C. Liu, S. Rawlins, and A. Imeson. Attributing physical and biological impacts to anthropogenic climate change. Nature, 453:353–357, 2008.

[54] C. Rosenzweig, D. Karoly, M. Vicarelli, P. Neofotis, Q. Wu, G. Casassa, A. Menzel, T. L. Root, N. Estrella, B. Seguin, P. Tryjanowski, C. Liu, S. Rawlins, and A. Imeson. Attributing physical and biological impacts to anthropogenic climate change. Nature, 453:353–357, 2008.

[55] T. Rotzer, R. Grote, and H. Pretzch. The timing of bud burst and its effect on tree growth. Int. J. Biometeorol., 48:109–118, 2004.

[56] P. Salembier and T. Sikora. Introduction to MPEG-7: Multimedia Content Descrip- tion Interface. John Wiley & Sons, Inc., New York, NY, USA, 2002.

[57] M. D. Schwartz. Phenology: an integrative environmental science. Kluwer Academic Publishers., 2003.

[58] M. D. Schwartz. Phenology: An Integrative Environmental Science. Academic Pu- blishers, 2003.

[59] M. D. Schwartz, B. C. Reed, and M. A. White. Assessing satellite derived start-of- season measures in the coterminous. International Journal of Climatology, 22:1793– 1805, 2002.

[60] D.C. Slaughter, D.K. Giles, and D. Downey. Autonomous robotic weed control sys- tems: A review. Computers and Electronics in Agriculture, 61(1):63 – 78, 2008. Emerging Technologies For Real-time and Integrated Agriculture Decisions.

[61] G. B. Souza and A. N. Marana. HTS and htsn: New shape descriptors based on hough transform statistics. Computer Vision and Image Understanding, 127:43–56, 2014.

[62] V. M. A. Souza, D. F. Silva, and G. E. A. P. A. Batista. Extracting texture features for time series classification. Stockholm, Sweden, 2014.

REFERÊNCIAS BIBLIOGRÁFICAS 51

[63] R. S. Torres and A. X. Falcão. Content-based image retrieval: Theory and applica- tions. RITA, 13(2):161–185, 2006.

[64] R. S. Torres, M. Hasegawa, S. Tabbone, J. Almeida, J. A. Santos, B. Alberton, and L. P. C. Morellato. Shape-based time series analysis for remote phenology studies. In 2013 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2013, Melbourne, Australia, July 21-26, 2013, pages 3598–3601, 2013.

[65] G. R. Walther. Plants in a warmer world. Perspectives in Plant Ecology Evolution and Systematics, 6:169–185, 2004.

[66] G. R. Walther. Plants in a warmer world. Perspectives in Plant Ecology Evolution and Systematics, 6:169–185, 2004.

[67] G. R. Walther, E. Post, P. Convey, A. Menzel, C. Parmesan, T. J. C. Beebee, J. M. Fromentin, O. H. Guldberg, and F. Bairlein. Ecological responses to recent climate change. Nature, 416:389–395, 2002.

[68] Q. Wang, J. Tenhunen, N. Q. Dinh, M. Reichstein, T. Vesala, and P. Keronen. Simi- larities in ground- and satellite-based {NDVI} time series and their relationship to physiological activity of a scots pine forest in finland. Remote Sensing of Environ- ment, 93(1–2):225 – 237, 2004.

[69] M. A. White, S. W. Running, and P. E. Thornton. The impact of growing-season length variability on carbon assimilation and evapotranspiration over 88 years in the eastern us deciduous forest. International Journal of Biometeorology, 42:139–145, 1999.

[70] L. Ye and E. J. Keogh. Time series shapelets: a new primitive for data mining. In Proceedings of the 15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Paris, France, June 28 - July 1, 2009, pages 947–956, 2009.

[71] D. Zhang and G. Lu. Review of shape representation and description. Pattern Recognition, 37(1):1–19, 2004.

[72] J. Zhao, Y. Zhang, Z. Tan, Q. Song, N. Liang, L. Yu, and J. Zhao. Using digital cameras for comparative phenological monitoring in an evergreen broad-leaved forest and a seasonal rain forest. Ecological Informatics, 10:65–72, 2012.