Trends in Ecology & Evolution
ReviewAnimal biometrics: quantifying and detecting phenotypic appearance
Section snippets
Animal biometrics: an emerging field that fingerprints phenotypic appearance
The field of animal biometrics (see Glossary) applies formal approaches to represent and detect phenotypic appearance. It can be used to recognize and classify species, identify individuals, detect the occurrence of, or variation in, a particular behavior, as well as to measure morphological traits and their interindividual variation or intraindividual changes over time (Box 1). Animal biometrics utilizes both the variability and uniqueness of coat patterns, vocalizations, movement dynamics,
What are animal biometrics? Computerized systems that recognize phenotypes
The source for data acquisition in animal biometrics is the measurable information displayed due to the anatomy or behavior of a species. Typically, aspects of the appearance of an animal, its movement characteristics, or vocalizations are selected and used as biometric entities [24]. Determining a suitable biometric entity set for an animal in a study population is a difficult task. The chosen traits must be measurable by a recording device, adequately permanent, characteristic of the animal
How do animal biometrics represent phenotypic appearance?
Representing and matching aspects of the phenotype in a quantifiable way is the central algorithmic challenge in animal biometrics. The key difficulty is how to capture the structural complexity exhibited by animal life using models and their parameters: animals actively change their shape and pose; animal surfaces reflect differently under different lighting; and animals frequently appear as partially hidden by other content, such as vegetation (Figure 1, Figure 2). Although computer graphic
How to use animal biometrics for profiling species and individuals
One of the most frequent applications of animal biometrics is the identification of individual animals. Similarly to the concept of minutiae in human fingerprints [27], uniqueness of animal appearance can typically be encoded by configurations of landmarks. Examples include elephant ear nicks [45], penguin spots [46], zebra stripe junctions [22], or SIFT features for Masai giraffe identification [47]. However, there is a trade-off between pattern variability versus constancy that typically
How to use animal biometrics for profiling behavior
The first automatic methods for behavioral phenotyping of captive mice [50] now exist, allowing for differentiation of behaviors including drinking, grooming, and resting. However, automated animal biometric applications that explicitly extract behaviors are essentially bound to controlled environments. Simplified approaches focus on analyzing footage in wildlife archives 23, 51, or consider less complex tasks, such as flock or group movements, as developed for bees [29], fish [52], or birds
Promising fields of application
Animal biometric systems have great potential for assisting in the filtering and indexing of audiovisual content that is increasingly produced in many ecological and evolutionary studies [24]. Handheld audiovisual devices, passive acoustic recording devices, and sensors carried by aerial vehicles are being increasingly used to document observations 57, 58, 59, 60. These routinely produce huge quantities of audiovisual data that are often at the limit of what can be processed manually. The time
Concluding remarks, recommendations, and outlook
The emerging field of animal biometrics is on the verge of providing powerful tools for field practitioners, ecologists, and researchers to use to collect and process phenotypic appearance information on species, individuals, their behavior, and morphology, in a standardized way and for a broad spectrum of applications. Although existing systems have shown that animal biometrics are feasible and useful to the biologist (Table 1); numerous challenges lie ahead to develop the field into a widely
Acknowledgments
This contribution is an output of SAISBECO, a joint project by Fraunhofer IDMT and IIS, Max Planck Institute for Evolutionary Anthropology, and the Visual Information Laboratory at the University of Bristol. We thank the providers of images (Kurt Amsler, Tobias Deschner, Karsten Dierks, Andreas Ernst, Jojo Head, Brad Norman, and Luisa Rabanal) and the Pan African Program. We are grateful to Doug Bolger and an anonymous reviewer for providing highly constructive feedback, and Peter Barham,
Glossary
- Algorithmic formalization
- description of an effective method by a finite list of well-defined instructions, such as a computer program.
- Animal biometrics
- quantified approaches for representing and detecting species, individuals, behaviors, or morphology traits based on measurable phenotypic characteristics.
- Annotation
- addition of information, such as species name, individual identity, pose information, lighting conditions, or object location, to audiovisual data for facilitating training of animal
References (74)
A survey on vision-based human action recognition
Image Vis. Comput.
(2010)Counting elusive animals: comparing field and genetic census of the entire mountain gorilla population of Bwindi impenetrable National Park, Uganda
Biol. Conserv.
(2009)Identifying elephant photos by multi-curve matching
Pattern Recognit.
(2008)A computer-vision system and methodology for the analysis of fish behaviour
Aquac. Eng.
(2012)Non-intrusive tracking of commercial broiler chickens in situ at different stocking densities
Appl. Anim. Behav. Sci.
(2008)The thin plate spline robust point matching (TPS-RPM) algorithm: a revisit
Pattern Recognit. Lett.
(2011)- et al.
Individuals and populations: the role of long-term, individual-based studies of animals in ecology and evolutionary biology
Trends Ecol. Evol.
(2010) Long-term datasets in biodiversity research and monitoring: assessing change in ecological communities through time
Trends Ecol. Evol.
(2010)A new dawn for citizen science
Trends Ecol. Evol.
(2009)Data-intensive science applied to broad-scale citizen science
Trends Ecol. Evol.
(2012)
Darwin's Camera: Art and Photography in the Theory of Evolution
Identification of individual Bewick's swans by bill patterns
Social organisation and behaviour of the Grevy's zebra
Z. Tierpsychol.
A test of computer-assisted matching using the North Pacific humpback whale, Megaptera novaeangliae, tail flukes photograph collection
Mar. Fish. Rev.
Object detection with discriminatively trained part based Models
IEEE Trans. Pattern Anal. Mach. Intell.
Outdoor human motion capture using inverse kinematics and von Mises-Fisher sampling
Real-time human pose recognition in parts from single depth images
Trainable convolution filters and their application to face recognition
IEEE Trans. Pattern Anal. Mach. Intell.
Learning and matching of dynamic shape manifolds for human action recognition
IEEE Trans. Image Process.
Differentiating spontaneous from posed facial expressions within a generic facial expression recognition framework
The use and abuse of photographic identification in sharks and rays
J. Fish Biol.
Spot the match: wildlife photo-identification using information theory
Front. Zool.
Heartbeat of a nest: using imagers as biological sensors
ACM Trans. Sens. Netw.
System and algorithms for an autonomous observatory assisting the search for the ivory-billed woodpecker
Estimating population size, structure, and residency time for whale sharks Rhincodon typus through collaborative photo-identification
Endangered Species Res.
Incorporating genotype uncertainty into mark recapture-type models for estimating abundance using DNA samples
Biometrics
Modeling misidentification errors in capture-recapture studies using photographic identification of evolving marks
Ecology
Estimating survival in photographic capture–recapture studies: overcoming misidentification error
Methods Ecol. Evol.
Computer-aided photograph matching in studies using individual identification: an example from Serengeti cheetahs
J. Mammal.
A tiger cannot change its stripes: using a three-dimensional model to match images of living tigers and tiger skins
Biol. Lett.
Real-time face detection and tracking of animals
Automated species identification: why not?
Philos. Trans. R. Soc. Lond. B
Biometric animal databases from field photographs: identification of individual zebra in the wild
Annual re-sightings of photographically identified white sharks (Carcharodon carcharias) at an eastern Pacific aggregation site (Guadalupe Island, Mexico)
Mar. Biol.
Identifying quadruped gait in wildlife video
Stereoscopic motion analysis in densely packed clusters: 3D analysis of the shimmering behaviour in giant honey bees
Front. Zool.
Cited by (150)
An experiment on animal re-identification from video
2023, Ecological InformaticsIdentification of free-ranging mugger crocodiles by applying deep learning methods on UAV imagery
2022, Ecological Informatics