Automatic Identification of Diatoms Using Decision Forests
A feature based identification scheme for microscopic images of diatoms is presented in this paper. Diatoms are unicellular algae found in water and other places wherever there is humidity and enough light for photo synthesis. The proposed automatic identification scheme follows a decision tree based classification approach. In this paper two different ensemble learning methods are evaluated and results are compared with those of single decision trees. As test sets two different diatom image databases are used. For each image in the databases general features like symmetry, geometric properties, moment invariants, and Fourier descriptors as well as diatom specific features like striae density and direction are computed.
KeywordsDecision Tree Recognition Rate Moment Invariant Fourier Descriptor Random Subspace
Unable to display preview. Download preview PDF.
- 1.Automatic Diatom Identification And Classification. Project home page: http://www.ualg.pt/adiac/.
- 4.S. Fischer, M. Binkert, and H. Bunke. Feature based retrieval of diatoms in an image database using decision trees. In ACIVS 2000, pages 67–72, Baden-Baden, Germany, August 2000.Google Scholar
- 5.S. Fischer, M. Binkert, and H. Bunke. Feature based retrieval of diatoms in an image database using decision trees. Technical Report IAM-00-001, Institute of Computer Science and Applied Mathematics, University of Bern, Switzerland, 2000.Google Scholar
- 6.S. Fischer, M. Binkert, and H. Bunke. Symmetry based indexing of diatoms in an image database.In Proceedings of the ICPR 2000, volume 2, pages 899–902, Barcelona, Spain, September 2000.Google Scholar
- 8.Y. Freund and R. E. Schapire. Experiments with a new boosting algorithm. In Machine Learning: Proceedings of the Thirteenth International Conference, pages 148–156, Bari, Italy, 1996. Morgan Kaufmann.Google Scholar
- 9.T. K. Ho. The random subspace method for constructing decision forests. IEEE Transactions on Pattern Analysis and Machine Intelligence, 20(8):832–844, August 1998.Google Scholar
- 10.M. Hu. Visual pattern recognition by moment invariants. IEEE Trans. Information Theory, 8(2):179–187, February 1962.Google Scholar
- 11.K. Krammer and H. Lange-Bertalot. Bacillariophyceae. In H. Ettl, J. Gerloff, H. Heynig, and D. Mollenhauer, editors, Süsswasserflora von Mitteleuropa (in German). Gustav Fischer Verlag, Stuttgart, 1986.Google Scholar
- 12.D. G. Mann, and S. J. M. Droop. Biodiversity, biogeography and conservation of diatoms. Hydrobiologia 336:19–32, 1996.Google Scholar
- 15.J. R. Quinlan. Bagging, boosting, and C4.5. In Proceedings of the Thirteenth National Conference on Artificial Intelligence, pages 725–730, Cambridge, MA, 1996. AAAI Press/MIT Press.Google Scholar
- 16.R. Quinlan. C4.5: Programs for Machine Learning. Morgan Kaufmann, San Mateo, USA, 1993.Google Scholar
- 17.P. L. Rosin. Measuring shape: Ellipticity, rectangularity, and triangularity. In Proceedings of the ICPR 2000, volume 2, pages 952–955, Barcelona, Spain, September 2000.Google Scholar
- 18.M. Sonka, V. Hlavac, and R. Boyle. Image Processing, Analysis, and Machine Vision. Brooks/Cole, 2. edition, 1999.Google Scholar
- 19.E. F. Stoermer, and T. B. Ladewski. Quantitative analysis of shape variations in type and modern populations of Gomphoneis herculeana. Nova Hedwigia Beihefte, 73:347–373, 1982.Google Scholar
- 20.E. F. Stoermer and J. P. Smol, editors. The Diatoms: Applications for the Environmental and Earth Science. Cambridge University Press, 1999.Google Scholar