Dataset Information

Camera settings and biome influence the accuracy of citizen science approaches to camera trap image classification.

ABSTRACT: Scientists are increasingly using volunteer efforts of citizen scientists to classify images captured by motion-activated trail cameras. The rising popularity of citizen science reflects its potential to engage the public in conservation science and accelerate processing of the large volume of images generated by trail cameras. While image classification accuracy by citizen scientists can vary across species, the influence of other factors on accuracy is poorly understood. Inaccuracy diminishes the value of citizen science derived data and prompts the need for specific best-practice protocols to decrease error. We compare the accuracy between three programs that use crowdsourced citizen scientists to process images online: Snapshot Serengeti, Wildwatch Kenya, and AmazonCam Tambopata. We hypothesized that habitat type and camera settings would influence accuracy. To evaluate these factors, each photograph was circulated to multiple volunteers. All volunteer classifications were aggregated to a single best answer for each photograph using a plurality algorithm. Subsequently, a subset of these images underwent expert review and were compared to the citizen scientist results. Classification errors were categorized by the nature of the error (e.g., false species or false empty), and reason for the false classification (e.g., misidentification). Our results show that Snapshot Serengeti had the highest accuracy (97.9%), followed by AmazonCam Tambopata (93.5%), then Wildwatch Kenya (83.4%). Error type was influenced by habitat, with false empty images more prevalent in open-grassy habitat (27%) compared to woodlands (10%). For medium to large animal surveys across all habitat types, our results suggest that to significantly improve accuracy in crowdsourced projects, researchers should use a trail camera set up protocol with a burst of three consecutive photographs, a short field of view, and determine camera sensitivity settings based on in situ testing. Accuracy level comparisons such as this study can improve reliability of future citizen science projects, and subsequently encourage the increased use of such data.

SUBMITTER: Egna N

PROVIDER: S-EPMC7663993 | biostudies-literature | 2020 Nov

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Camera settings and biome influence the accuracy of citizen science approaches to camera trap image classification.

Egna Nicole N O'Connor David D Stacy-Dawes Jenna J Tobler Mathias W MW Pilfold Nicholas N Neilson Kristin K Simmons Brooke B Davis Elizabeth Oneita EO Bowler Mark M Fennessy Julian J Glikman Jenny Anne JA Larpei Lexson L Lekalgitele Jesus J Lekupanai Ruth R Lekushan Johnson J Lemingani Lekuran L Lemirgishan Joseph J Lenaipa Daniel D Lenyakopiro Jonathan J Lesipiti Ranis Lenalakiti RL Lororua Masenge M Muneza Arthur A Rabhayo Sebastian S Ole Ranah Symon Masiaine SM Ruppert Kirstie K Owen Megan M

Ecology and evolution 20201006 21

Scientists are increasingly using volunteer efforts of citizen scientists to classify images captured by motion-activated trail cameras. The rising popularity of citizen science reflects its potential to engage the public in conservation science and accelerate processing of the large volume of images generated by trail cameras. While image classification accuracy by citizen scientists can vary across species, the influence of other factors on accuracy is poorly understood. Inaccuracy diminishes ...[more]

PMID: 33209262

Similar Datasets

Project description:BackgroundCitizen science approaches, which involve members of the public as active collaborators in scientific research, are increasingly being recognized for their potential benefits in chronic disease prevention. However, understanding the potential applicability, feasibility and impacts of these approaches is necessary if they are to be more widely used. This study aimed to synthesize research that has applied and evaluated citizen science approaches in chronic disease prevention and identify key questions, gaps, and opportunities to inform future work in this field.MethodsWe searched six databases (Scopus, Medline, Embase, PsycInfo, PubMed, and CINAHL) in January 2022 to identify articles on the use of citizen science in prevention. We extracted and synthesized data on key characteristics of citizen science projects, including topics, aims and level of involvement of citizen scientists, as well as methods and findings of evaluations of these projects.ResultsEighty-one articles reported on citizen science across a variety of health issues, predominantly physical activity and/or nutrition. Projects primarily aimed to identify problems from the perspective of community members; generate and prioritize solutions; develop, test or evaluate interventions; or build community capacity. Most projects were small-scale, and few were co-produced with policy or practice stakeholders. While around half of projects included an evaluation component, overall, there was a lack of robust, in-depth evaluations of the processes and impacts of citizen science projects.ConclusionsCitizen science approaches are increasingly being used in chronic disease prevention to identify and prioritize community-focused solutions, mobilize support and advocacy, and empower communities to take action to support their health and wellbeing. However, to realize the potential of this approach more attention needs to be paid to demonstrating the feasibility of using citizen science approaches at scale, and to rigorous evaluation of impacts from using these approaches for the diverse stakeholders involved.

Dataset Information

Camera settings and biome influence the accuracy of citizen science approaches to camera trap image classification.

Publications

Camera settings and biome influence the accuracy of citizen science approaches to camera trap image classification.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets