Biodiversity Databases: Techniques, Politics, and Applications - Chapter 5 ppsx

Morphometric and Artificial Neural Network Approaches to the Automated Species Recognition Problem in Systematics Norman MacLeod, M.. Status within the Systematics Community?...58 5.2..

Morphometric and Artificial Neural Network Approaches to the Automated Species Recognition

Problem in Systematics

Norman MacLeod, M O’Neill and Steven A Walsh

Contents

Abstract 37

5.1 Introduction 38

5.1.1 The.Need.for.Automated.Species.Recognition.in.Systematics 38

5.1.2 Approaches 40

5.1.3 Objectives 43

5.1.4 Materials.and.Methods 43

5.1.5 Results 47

5.1.6 Discussion 53

5.1.6.1 Which.Approach? 53

5.1.6.2 Scope for Synthesis? 57

5.1.6.3 Further Research Directions? 57

5.1.6.4 Status within the Systematics Community? 58

5.2 Summary.and.Conclusions 60

Acknowledgements 61

References 61

AbstrACt

One.approach.to.addressing.long-standing.concerns.associated.with.the.taxonomic.imped-iment and occasional low reproducibility of taxonomic data is through development of

automated species identification systems Such systems can, in principle, be combined

with.image-based.or.image-.and.text-based.taxonomic.databases.to.add.elements.of.expert

system.functionality Two.generalized.approaches.are.considered.relevant.in.this.context:

morphometric systems based on some form of linear discriminant analysis (LDA) and

economic imperatives are rapidly drawing to a close In order to attract personnel and.

resources, morphology-based taxonomy must transform itself into a ‘large, coordinated,

international.scientific.enterprise’.(Wheeler,.2003,.p 4) Many.have.recently.touted.use.of

they.are.used.in.a.wide.variety.of.scientific.contexts.(e.g.,.oceanography,.biogeog-logical.diversity;.and

of within-group ordinations within the shape spaces defined by CVA axes The

digital.automated.image-analysis.system.(DAISY;.Weeks.et.al 1997,.1999a,.b) This.imple-mentation accepts training sets in the form of standard format images (e.g., jpeg, tiff)

of authoritatively identified specimens These image-based training sets were processed

comparison between brightness values between pixel locations The result allows each.

Bookstein 1990; Marcus et al 1993, 1996; MacLeod and Forey 2002) These variables.

Once again, using least-squares superposition to normalize the coordinate data for

generalized size differences (thereby achieving an entirely shape-based discrimination)

and.employing.CVA.to.construct.a.discriminant.space,.an.unprecedented.correct.cross-validation.identification.ratio.of.0.99.was.obtained.(Table.5.1.and.Figure.5.8) Of.the.two

misidentified.specimens, a.Globigerinelloides conglobatus.was.mistaken.completely for.

Globigerinelloides ruber (posterior.probability.=.1.00).while.a.Globigerinella

inaequilate-ralis was.ambiguously.mistaken.for.Globorotalia tumida.(posterior.probability.=.0.67).

Cross-validation.results.for.the.DAISY-based.ANN.analysis.differ.from.those.of.the

CVA.analysis.in.terms.of.the.manner.in.which.the.posterior.probabilities.are.calculated

Instead of using a distance-based approach for assigning unknowns to groups, DAISY

uses.a.combined eightfold.distance-coordination approach.with the.minimum

frame This approach relaxes the morphometric requirement for landmarks to represent.

a comparatively small number but biologically well-known set of close topological

cor-respondences between objects in favour of more inclusive information drawn from the

both approaches are limited by complimentary deficiencies: morphometric methods are

rich in biological meaning, but deficient in overall geometric information content while

ANN.methods.are.rich.in.overall.information.content,.but.deficient.in.biological.meaning

A.synthesis.between.the.two.is.not.only.possible,.but.highly.desirable

Until such a synthesis is achieved, however, it makes sense to match the available

strengths of each approach to the diversity of morphological problems at hand

Mor-phometrics would appear to be best utilized for the investigation of precise distinctions

...   Further Research Directions?

For morphometric and ANN approaches, one of the most important needs is for better

specification of adequate training set attributes In the technical literature produced on

these.methods.over.the.years,.scarcely.any.but.the.most.general.statements.about.the.com-position.and.nature.of.reliable.training.sets.have.been.made To.be.sure,.a.large.body.of

concerns Systems that can authoritatively achieve consistent, semi-automated and fully

automated identifications of planktonic microfossil species — and, by extension, many

A.major.factor.hold-to.perform.to.necessary taxonomic identifications to.a.high degree of.accuracy

systematic.research.guiding.the.discovery.and.testing.of.new.characters.and.refin-Reinvigoration of the discipline of morphological systematiclenges.such.as.DNA.bar.coding.and.GeneBank,.morphological.systematics.must

s In.the.face.of.chal-ibly.morphology-based.refugia.(e.g.,.palaeontology) Because.of.their.generality,

become.more.automated.and.efficient.or.it.will.cease.to.exist.outside.a.few.irreduc-automated species (= image) recognition systems can be used in a wide variety

•

•

•

•

sis This ability extends across the spectrum of systematic data (e.g., morphol-ogy,.ecology,.geography,.stratigraphic,.chemical,.molecular,.audio,.olfactory,.DNA.

MacLeod,.N (1998).Impacts.and.marine.invertebrate.extinctions In.Meteorites: Flux with time and

impact effects,.ed N.M Grady,.R Hutchinson,.G.J.H McCall,.and.D.A Rotherby Geological

Society.of.London,.London,.217–246

MacLeod,.N (1999).Generalizing.and.extending.the.eigenshape.method.of.shape.visualization.and

analysis Paleobiology.25(1):.107–138.

Ripley,.B.D (1996).Pattern recognition and neural networks Cambridge.University.Press,.Cambridge.

Rohlf,.F.J and.Bookstein,.F.L (1990).Proceedings of the Michigan morphometrics workshop The.