Current Science

Open Access Open Access  Restricted Access Subscription Access

Entropy Query by Bagging-Based Active Learning Approach in the Extreme Learning Machine Framework for Hyperspectral Image Classification


Affiliations
1 School of Computer and Systems Sciences, Jawaharlal Nehru University, New Delhi 110 067, India
2 School of Electronics Engineering, Kalinga Institute of Industrial Technology, Bhubaneswar 751 024, India
 

Active learning (AL) technique is the classification of remote sensing images, where collecting efficient training data is costly in terms of labour and the time taken. The prime objective of AL technique is to obtain high classification accuracy with the training sample as compact as possible. Most studies on the classification of remote sensing images using AL, focused only on accuracy, with hardly any study on computation time. Keeping reduction of computation time as the objective, here we present, an entropy query by bagging (EQB)-based AL approach in the extreme learning machine (ELM) framework for remote sensing image classification. The performance of this approach is compared with the widely used support vector machine (SVM) AL framework in combination with different query strategies. To verify the efficacy of the study, the approaches were tested on two hyperspectral remote-sensing images, i.e. Kennedy Space Centre (KSC) and Botswana (BOT). The proposed system depicts competitive classification performance while significantly reducing computation time.

Keywords

Active Learning, Computation Time, Extreme Learning Machine, Entropy Query by Bagging.
User
Notifications
Font Size

  • Pearlman, J. S., Berry, P. S., Segal, C. C., Shapanski, J., Beiso, D. and Carman, S. L., Hyperion: a space-based imaging spectrometer. IEEE Trans. Geosci. Remote Sensing, 2003, 41(6), 1160–1173.

  • Heiden, U., Heldens, W., Roessner, S., Segl, K., Esch, T. and Mueller, A., Urban structure type characterization using hyperspectral remote sensing and height information. Lands. Urban Plann., 2012, 105(4), 361–375.

  • Giri, C., Observation and monitoring of mangrove forests using remote sensing: opportunities and challenges. Remote Sensing, 2016, 8(9), 1–8.

  • Sahoo, R. N., Ray, S. S. and Manjunath, K. R., Hyperspectral remote sensing of agriculture. Curr. Sci., 2015, 108(5), 848–859.

  • Jiang, J. and Tian, G., Analysis of the impact of land use/land cover change on land surface temperature with remote sensing. Proc. Environ. Sci., 2010, 2, 571–575.

  • Fauvel, M., Tarabalka, Y., Benediktsson, J. A., Chanussot, J. and Tilton, J. C., Advances in spectral–spatial classification of hyperspectral images. Proc. IEEE, 2013, 101(3), 652–675.

  • Camps-Valls, G., Tuia, D., Bruzzone, L. and Benediktsson, J. A., Advances in hyperspectral image classification: earth monitoring with statistical learning methods. IEEE Signal Process. Mag., 2014, 31(1), 45–54.

  • Kuo, B., Ho, H., Li, C., Hung, C. and Taur, J., A kernel-based feature selection method for SVM with RBF Kernel for hyperspectral image classification. IEEE J. Sel. Top. Appl. Earth Observ. Remote Sensing, 2014, 7(1), 317–322.

  • Fan, Y., Zheng, S., Yang, H., Zhang, C. and Su, H., Causalityweighted active learning for abnormal event identification-based on the topic model. Opt. Eng., 2012, 51(7), 077204-1–077204-12.

  • Han, Y., Li, P., Zhang, Y., Hong, Z., Liu, K. and Wang, J., Combining active learning and transductive support vector machines for sea ice detection. J. Appl. Remote Sensing, 2018, 12(2), 026016.

  • Di, W. and Crawford, M. M., View generation for multiview maximum disagreement based active learning for hyperspectral image classification. IEEE Trans. Geosci. Remote Sensing, 2012, 50(5), 1942–1954.

  • Mitra, P., Murthy, C. A. and Pal, S. K., A probabilistic active support vector learning algorithm. IEEE Trans. Pattern Anal. Mach. Intel., 2004, 26(3), 413–418.

  • Rajan, S., Ghosh, J. and Crawford, M. M., An active learning approach to hyperspectral data classification. IEEE Trans. Geosci. Remote Sensing, 2008, 46(4), 1231–1242.

  • Tuia, D., Pasolli, E. and Emery, W. J., Using active learning to adapt remote sensing image classifiers. Remote Sensing Environ., 2011, 115(9), 2232–2242.

  • Crawford, M. M., Tuia, D. and Yang, H. L., Active learning: any value for classification of remotely sensed data. Proc. IEEE, 2013, 101(3), 593–608.

  • Pradhan, M. K., Minz, S. and Shrivastava, V. K., A kernel-based extreme learning machine framework for classification of hyperspectral images using active learning. J. Indian Soc. Remote Sensing, 2019, 47(10), 1693–1705.

  • Jamshidpour, N., Safari, A. and Homayouni, S., Multiview active learning optimization based on genetic algorithm and Gaussian mixture models for hyperspectral data. IEEE Geosci. Remote Sensing Lett., 2019, 17(1), 172–176.

  • Pradhan, M. K., Minz, S. and Shrivastava, V. K., Fisher discriminant ratio based multiview active learning for the classification of remote sensing images. In 4th International Conference on Recent Advances in Information Technology, Dhanbad, India, 2018, pp. 1–6.

  • Tuia, D., Volpi, M., Copa, L., Kanevski, M. and Munoz-Mari, J., A survey of active learning algorithms for supervised remote sensing image classification. IEEE J. Sel. Top. Signal Process., 2011, 5(3), 606–617.

  • Tuia, D., Ratle, F., Pacifici, F., Kanevski, M. F. and Emery, W. J., Active learning methods for remote sensing image classification. IEEE Trans. Geosci. Remote Sensing, 2009, 47(7), 2218–2232.

  • Scheffer, T., Decomain, C. and Wrobel, S., Active hidden Markov models for information extraction. In Proceedings of the 4th International Symposium on Intelligent Data Analysis, Cascais, Portugal, 2001, pp. 309–318.

  • Pasolli, E., Melgani, F. and Bazi, Y., Support vector machine active learning through significance space construction. IEEE Geosci. Remote Sensing Lett., 2011, 8(3), 431–435.

  • Demir, B., Persello, C. and Bruzzone, L., Batch-mode activelearning methods for the interactive classification of remote sensing images. IEEE Trans. Geosci. Remote Sensing, 2011, 49(3), 1014– 1031.

  • Li, J., Bioucas-Dias, J. M. and Plaza, A., Semisupervised hyperspectral image segmentation using multinomial logistic regression with active learning. IEEE Trans. Geosci. Remote Sensing, 2010, 48(11), 4085–4098.

  • Li, J., Bioucas-Dias, J. M. and Plaza, A., Hyperspectral image segmentation using a new Bayesian approach with active learning. IEEE Trans. Geosci. Remote Sensing, 2011, 49(10), 3947–3960.

  • Platt, J. C., Probabilistic outputs for support vector machines and comparisions to regulized likelihood methods. In Adv. Large Margin Classifiers, 1999, 10(3), 61–74.

  • MacKay, D. J. C., Information-based objective functions for active data selection. Neural Comput., 1992, 4, 590–604.

  • Luo, T., Kramer, K., Goldgof, D. B., Hall, L. O., Samson, S., Remsen, A. and Hopkins, T., Active learning to recognize multiple types of plankton. J. Mach. Learn. Res., 2005, 6, 589–613.

  • Zhou, Y. and Goldman, S., Democratic co-learning. In 16th IEEE International Conference, Tools with Artificial Intelligence, Boca Raton, FL, USA, 2004, pp. 594–602.

  • Abe, N. and Mamitsuka, H., Query learning strategies using boosting and bagging. In Proc. ICML, Madison, WI, USA, 1998, pp. 1–9.

  • Melville, P. and Mooney, R., Diverse ensembles for active learning. In Proceedings of the 21st International Conference on Machine Learning, Banff, Canada, 2004, pp. 584–591.

  • Xu, X., Li, J. and Li, S., Multiview intensity-based active learning for hyperspectral image classification. IEEE Trans. Geosci. Remote Sensing, 2018, 56(2), 669–680.

  • Wan, L., Tang, K., Li, M., Zhong, Y. and Qin, A. K., Collaborative active and semisupervised learning for hyperspectral remote sensing image classification. IEEE Trans. Geosci. Remote Sensing, 2015, 53(5), 2386–2396.

  • Patra, S., Bhardwaj, K. and Bruzzone, L., A spectral–spatial multicriteria active learning technique for hyperspectral image classification. IEEE J. Sel. Top. Appl. Earth Observ. Remote Sensing, 2017, 10(12), 5213–5227.

  • Huang, G. B., Zhu, Q. Y. and Siew, C. K., Extreme learning machine: theory and applications. Neurocomputing, 2006, 70(1–3), 489–501.

  • Huang, G. B., Wang, D. H. and Lan, Y., Extreme learning machine: a survey. Int. J. Mach. Learn. Cybern., 2011, 2(2), 107–122.

  • Huang, G. B., Zhou, H., Ding, X. and Zhang, R., Extreme learning machine for regression and multiclass classification. IEEE Trans. Syst., Man Cybern., 2012, 42(2), 513–529.

  • Pal, M., Maxwell, A. E. and Warner, T. A., Kernel-based extreme learning machine for remote-sensing image classification. Remote Sensing Lett., 2013, 4(9), 853–862.

  • Bazi, Y., Alajlan, N., Melgani, F., AlHichri, H., Malek, S. and Yager, R. R., Differential evolution extreme learning machine for the classification of hyperspectral images. IEEE Geosci. Remote Sensing Lett., 2014, 11(6), 1066–1070.

  • Samat, A., Du, P., Liu, S. , Li, J. and Cheng, L., E2LMs: ensemble extreme learning machines for hyperspectral image classification. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sensing, 2014, 7(4), 1060–1069.

  • Heras, D. B., Argüello, F. and Quesada-Barriuso, P., Exploring ELM based spatial–spectral classification of hyperspectral images. Int. J. Remote Sensing, 2014, 35(2), 401–423.

  • Pradhan, M. K., Minz, S. and Shrivastava, V. K., Fast active learning for hyperspectral image classification using extreme learning machine. IET Image Process., 2018, 13(4), 549–555.

  • Li, M. and Sethi, I. K., Confidence-based active learning. IEEE Trans. Pattern Anal. Mach. Intellige., 2006, 28(8), 1251–1261.

  • Seung, H. S., Opper, M. and Sompolinsky, H., Query by committee. In Proceedings of the Fifth Annual Workshop Computational Learning Theory, Pittsburgh, PA, USA, 1992, pp. 287–294.

  • Freund, Y., Seung, H. S., Shamir, E. and Tishby, N., Selective sampling using the query by committee algorithm. Mach. Learn., 1997, 28(2–3), 133–168.

  • Efron, B., Bootstrap methods: another look at the jackknife. In Breakthroughs Statistics, Springer, New York, USA, 1992, pp. 569–593.

  • HSI dataset: KSC and BOT; http://www.ehu.eus/ccwintco/index.php/Hyperspectral_Remote_Sensing_Scenes (accessed on 22 September 2017).

  • Chang, C. C. and Lin, C. J., LIBSVM: a library for support vector machines. ACM Trans. Intel. Syst. Techn., 2011, 2(3), 27; http://www.csie.ntu.edu.tw/~cjlin/libsvm (accessed on 22 October 2017).

  • Wang, Z., Du, B., Zhang, L., Zhang, L. and Jia, X., A novel semisupervised active-learning algorithm for hyperspectral image classification. IEEE Trans. Geosci. Remote Sensing, 2017, 55(6), 3071–3083.


Abstract Views: 400

PDF Views: 124




  • Entropy Query by Bagging-Based Active Learning Approach in the Extreme Learning Machine Framework for Hyperspectral Image Classification

Abstract Views: 400  |  PDF Views: 124

Authors

Monoj K. Pradhan
School of Computer and Systems Sciences, Jawaharlal Nehru University, New Delhi 110 067, India
Sonajharia Minz
School of Computer and Systems Sciences, Jawaharlal Nehru University, New Delhi 110 067, India
Vimal K. Shrivastava
School of Electronics Engineering, Kalinga Institute of Industrial Technology, Bhubaneswar 751 024, India

Abstract


Active learning (AL) technique is the classification of remote sensing images, where collecting efficient training data is costly in terms of labour and the time taken. The prime objective of AL technique is to obtain high classification accuracy with the training sample as compact as possible. Most studies on the classification of remote sensing images using AL, focused only on accuracy, with hardly any study on computation time. Keeping reduction of computation time as the objective, here we present, an entropy query by bagging (EQB)-based AL approach in the extreme learning machine (ELM) framework for remote sensing image classification. The performance of this approach is compared with the widely used support vector machine (SVM) AL framework in combination with different query strategies. To verify the efficacy of the study, the approaches were tested on two hyperspectral remote-sensing images, i.e. Kennedy Space Centre (KSC) and Botswana (BOT). The proposed system depicts competitive classification performance while significantly reducing computation time.

Keywords


Active Learning, Computation Time, Extreme Learning Machine, Entropy Query by Bagging.

References





DOI: https://doi.org/10.18520/cs%2Fv119%2Fi6%2F934-943