Comparison between automated analysis of zooplankton using ZooImage and traditional methodology

The objective of this study was to evaluate the accuracy of the ZooImage image analysis system for taxonomic classification of zooplankton samples. For this purpose, automated analysis with the ZooImage software was compared with traditional analysis, using zooplankton samples collected in the Iceland Sea in July 2006. When compared with the traditional methodology, ZooImage was able to classify zooplankton into main taxonomic entities (size classes and families or genera in some cases), while being less successful in identifying the zooplankton into species. Other important information, that is difficult and time consuming to obtain by traditional methods such as biomass and size distributions are, however, easily obtained with ZooImage. The automated analysis takes much less time than the traditional methods. While the study confirms that ZooImage is a promising tool for rapidly analysing zooplankton samples, it is also clear that the traditional approach will be needed in future investigations, particularly studies addressing zooplankton community structure, development and life history.

[1]  Philippe Grosjean,et al.  Enumeration, measurement, and identification of net zooplankton samples using the ZOOSCAN digital imaging system , 2004 .

[2]  A.Z. Kouzani,et al.  Face classification by a random forest , 2007, TENCON 2007 - 2007 IEEE Region 10 Conference.

[3]  Einar Hjörleifsson,et al.  MacImage: a user-friendly image-analysis system for the accurate mensuration of marine organisms , 1986 .

[4]  Mar Ecol Ser Prog,et al.  Biological pattern recognition by neural networks , .

[5]  P. Utgoff,et al.  RAPID: Research on Automated Plankton Identification , 2007 .

[6]  Tony R. Martinez,et al.  Decision Tree Ensemble: Small Heterogeneous Is Better Than Large Homogeneous , 2008, 2008 Seventh International Conference on Machine Learning and Applications.

[7]  J. A. Lozano,et al.  Optimizing the number of classes in automated zooplankton classification , 2009 .

[8]  R. Hopcroft,et al.  Assessment of ZooImage as a tool for the classification of zooplankton , 2008 .

[9]  Qiao Hu,et al.  Accurate automatic quantification of taxa-specific plankton abundance using dual classification with correction , 2006 .

[10]  Philippe Grosjean,et al.  Spring zooplankton distribution in the Bay of Biscay from 1998 to 2006 in relation with anchovy recruitment , 2008 .

[11]  Steve Horvath,et al.  Tumor classification by tissue microarray profiling: random forest clustering applied to renal cell carcinoma , 2005, Modern Pathology.

[12]  Manfred Rolke,et al.  Size structure analysis of zooplankton samples by means of an automated image analyzing system , 1984 .

[13]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[14]  Phil F. Culverhouse,et al.  Biological pattern recognition by neural networks , 1991 .

[15]  H. Edgerton,et al.  Silhouette photography of oceanic zooplankton , 1979, Nature.

[16]  C. Davis,et al.  Real-time observation of taxa-specific plankton distributions: an optical sampling method , 2004 .

[17]  S. Motoda.  Devices of Simple Plankton Apparatus , 1959 .

[18]  M. Alcaraz,et al.  Estimating zooplankton biomass through image analysis , 2003 .

[19]  X. Irigoien,et al.  Mapping plankton distribution in the Bay of Biscay during three consecutive spring surveys , 2007 .

[20]  P. Culverhouse,et al.  Do experts make mistakes? A comparison of human and machine identification of dinoflagellates , 2003 .