Deep learning-based appearance features extraction for automated carp species identification

Animal biometrics is an emerging field that develops quantified approaches for representing and detecting the phenotypic appearance of species, individuals, behaviors, and morphological traits. It operates at the intersection between pattern recognition, ecology, and information sciences, producing computerized systems for phenotypic measurement and interpretation. Animal biometrics can benefit a wide range of disciplines, including biogeography, population ecology, and behavioral research. Currently, real-world applications are gaining momentum, augmenting the quantity and quality of ecological data collection and processing. However, to advance animal biometrics will require integration of methodologies among the scientific disciplines involved. Such efforts will be worthwhile because the great potential of this approach rests with the formal abstraction of phenomics, to create tractable interfaces between different organizational levels of life.

Animal Biometrics: quantifying and detecting phenotypic appearance

Due to the ongoing biodiversity crisis, many species including great apes like chimpanzees are on the brink of extinction. Consequently, there is an urgent need to protect the remaining populations of threatened species. To overcome the catastrophic decline of biodiversity, biologists and gamekeepers recently started to use remote cameras and recording devices for wildlife monitoring in order to estimate the size of remaining populations. However, the manual analysis of the resulting image and video material is extremely tedious, time consuming, and cost intensive. To overcome the burden of time-consuming routine work, we have recently started to develop computer vision algorithms for automated chimpanzee detection and identification of individuals. Based on the assumption that humans and great apes share similar properties of the face, we proposed to adapt and extend face detection and recognition algorithms, originally developed to recognize humans, for chimpanzee identification. In this paper we do not only summarize our earlier work in the field, we also extend our previous approaches towards a more robust system which is less prone to difficult lighting situations, various poses, and expressions as well as partial occlusion by branches, leafs, or other individuals. To overcome the limitations of our previous work, we combine holistic global features and locally extracted descriptors using a decision fusion scheme. We present an automated framework for photo identification of chimpanzees including face detection, face alignment, and face recognition. We thoroughly evaluate our proposed algorithms on two datasets of captive and free-living chimpanzee individuals which were annotated by experts. In three experiments we show that the presented framework outperforms previous approaches in the field of great ape identification and achieves promising results. Therefore, our system can be used by biologists, researchers, and gamekeepers to estimate population sizes faster and more precisely than the current frameworks. Thus, the proposed framework for chimpanzee identification has the potential to open up new venues in efficient wildlife monitoring and can help researches to develop innovative protection schemes in the future.

/pdf/an-automated-chimpanzee-identification-system-using-face-2xm5p5mkrj.pdf

An automated chimpanzee identification system using face detection and recognition

In this paper we report on the context and evaluation of a system for an automatic interpretation of sightings of individual western lowland gorillas (Gorilla gorilla gorilla) as captured in facial field photography in the wild. This effort aligns with a growing need for effective and integrated monitoring approaches for assessing the status of biodiversity at high spatio-temporal scales. Manual field photography and the utilisation of autonomous camera traps have already transformed the way ecological surveys are conducted. In principle, many environments can now be monitored continuously, and with a higher spatio-temporal resolution than ever before. Yet, the manual effort required to process photographic data to derive relevant information delimits any large scale application of this methodology. The described system applies existing computer vision techniques including deep convolutional neural networks to cover the tasks of detection and localisation, as well as individual identification of gorillas in a practically relevant setup. We evaluate the approach on a relatively large and challenging data corpus of 12,765 field images of 147 individual gorillas with image-level labels (i.e. missing bounding boxes) photographed at Mbeli Bai at the Nouabal-Ndoki National Park, Republic of Congo. Results indicate a facial detection rate of 90.8% AP and an individual identification accuracy for ranking within the Top 5 set of 80.3%. We conclude that, whilst keeping the human in the loop is critical, this result is practically relevant as it exemplifies model transferability and has the potential to assist manual identification efforts. We argue further that there is significant need towards integrating computer vision deeper into ecological sampling methodologies and field practice to move the discipline forward and open up new research horizons.

/pdf/towards-automated-visual-monitoring-of-individual-gorillas-at5swuvlii.pdf

Towards Automated Visual Monitoring of Individual Gorillas in the Wild

Wildlife managers are urgently searching for improved sociodemographic population assessment methods to evaluate the effectiveness of implemented conservation activities. These need to be inexpensive, appropriate for a wide spectrum of species and straightforward to apply by local staff members with minimal training. Furthermore, conservation management would benefit from single approaches which cover many aspects of population assessment beyond only density estimates, to include for instance social and demographic structure, movement patterns, or species interactions. Remote camera traps have traditionally been used to measure species richness. Currently, there is a rapid move toward using remote camera trapping in density estimation, community ecology, and conservation management. Here, we demonstrate such comprehensive population assessment by linking remote video trapping, spatially explicit capture–recapture (SECR) techniques, and other methods. We apply it to three species: chimpanzees Pan troglodytes troglodytes, gorillas Gorilla gorilla gorilla, and forest elephants Loxodonta cyclotis in Loango National Park, Gabon. All three species exhibited considerable heterogeneity in capture probability at the sex or group level and density was estimated at 1.72, 1.2, and 1.37 individuals per km2 and male to female sex ratios were 1:2.1, 1:3.2, and 1:2 for chimpanzees, gorillas, and elephants, respectively. Association patterns revealed four, eight, and 18 independent social groups of chimpanzees, gorillas, and elephants, respectively: key information for both conservation management and studies on the species' ecology. Additionally, there was evidence of resident and nonresident elephants within the study area and intersexual variation in home range size among elephants but not chimpanzees. Our study highlights the potential of combining camera trapping and SECR methods in conducting detailed population assessments that go far beyond documenting species diversity patterns or estimating single species population size. Our study design is widely applicable to other species and spatial scales, and moderately trained staff members can collect and process the required data. Furthermore, assessments using the same method can be extended to include several other ecological, behavioral, and demographic aspects: fission and fusion dynamics and intergroup transfers, birth and mortality rates, species interactions, and ranging patterns.

https://onlinelibrary.wiley.com/doi/pdf/10.1002/ece3.670

Effective sociodemographic population assessment of elusive species in ecology and conservation management

In recent years, thousands of species populations declined catastrophically leaving many species on the brink of extinction. Several biological studies have shown that especially primates like chimpanzees and gorillas are threatened. An essential part of effective biodiversity conservation management is population monitoring using remote camera devices. However, due to the large amount of data, the manual analysis of video recordings is extremely time consuming and highly cost intensive. Consequently, there is a high demand for automatic analytical routine procedures using computer vision techniques to overcome this issue. In this paper we present a technique for the identification of great apes, in particular chimpanzees, using state-of-the-art algorithms for human face recognition in combination with several classification schemes. For benchmark purposes we provide a publicly available dataset of captive chimpanzees. In our experiments we applied several common techniques like the well known Eigenfaces, Fisherfaces, Laplacianfaces and Randomfaces approaches to identify individuals. We compare all of these methods in combination with the classification approaches Nearest Neighbor (NN), Support Vector Machine (SVM) and a new concept for face recognition, Sparse Representation Classification (SRC) based on Compressive Sensing (CS).

Identification of great apes using face recognition

In this paper we propose and evaluate a face recognition approach for the individual identification of great apes. We extend our previous work to a more automatic approach using an unsupervised face alignment method known as congealing instead of a projective transform based on manually annotated facial feature points. Furthermore we present an improved version of the Randomfaces approach, called Hybridfaces, which complements the global recognition results with information obtained from local facial regions. We evaluate our approach on three publicly available primate databases of captive chimpanzees, free-living chimpanzees, and free-living western lowland gorillas. Our proposed framework shows promising results and the Hybridfaces approach clearly outperforms the previously used basic Randomfaces method on all three datasets.

Alexander Loos

Papers

Identification of great apes using face recognition

Towards automated visual identification of primates using face recognition