Economical crowdsourcing for camera trap image classification

Abstract

Camera trapping is widely used to monitor mammalian wildlife but creates large image datasets that must be classified. In response, there is a trend towards crowdsourcing image classification. For high-profile studies of charismatic faunas, many classifications can be obtained per image, enabling consensus assessments of the image contents. For more local-scale or less charismatic communities, however, demand may outstrip the supply of crowdsourced classifications. Here, we consider MammalWeb, a local-scale project in North East England, which involves citizen scientists in both the capture and classification of sequences of camera trap images. We show that, for our global pool of image sequences, the probability of correct classification exceeds 99% with about nine concordant crowdsourced classifications per sequence. However, there is high variation among species. For highly recognizable species, species-specific consensus algorithms could be even more efficient; for difficult to spot or easily confused taxa, expert classifications might be preferable. We show that two types of incorrect classifications -misidentification of species and overlooking the presence of animals -have different impacts on the confidence of consensus classifications, depending on the true species pictured. Our results have implications for data capture and classification in increasingly numerous, local-scale citizen science projects. The species-specific nature of our findings suggests that the performance of crowdsourcing projects is likely to be highly sensitive to the local fauna and context. The generality of consensus algorithms will, thus, be an important consideration for ecologists interested in harnessing the power of the crowd to assist with camera trapping studies.

Wild boar in focus: initial model outputs of wild boar distribution based on occurrence data and identification of priority areas for data collection

“…Such sighting data could be collected with the help of citizen science recording projects (Hsing et al, 2018). In the absence of data on true absence or on actual survey effort this makes a strong case for expanding the WBDM model to collect occurrence data of other artiodactyl species, many of which are game species and should be recorded by similar organisations as wild boar.…”

“…In the absence of data on true absence or on actual survey effort this makes a strong case for expanding the WBDM model to collect occurrence data of other artiodactyl species, many of which are game species and should be recorded by similar organisations as wild boar. Such sighting data could be collected with the help of citizen science recording projects (Hsing et al, 2018). The present document has been produced and adopted by the bodies identified above as authors.…”

Adopting Citizen Science as a Tool to Enhance Monitoring for an Environment Agency

“…Remote camera images are used to monitor the health of threatened species populations or to monitor trends in abundance of species. Involving citizen scientists can be a popular means of engaging people to contribute to the development of long-term and frequent observation datasets (Swanson et al 2015;Hsing et al 2018). However, uncertainty around the value of public provision of images and participation in curation remained.…”

Wildlife surveillance using deep learning methods

“…The conventional approach of sifting through images by eye can be laborious and expensive (although some studies have reduced costs by crowd sourcing; e.g., Hsing et al (2018)). The conventional approach of sifting through images by eye can be laborious and expensive (although some studies have reduced costs by crowd sourcing; e.g., Hsing et al (2018)).…”

“…This surge of interest in remote surveillance has, however, been accompanied by increasing recognition of the challenges associated with screening the enormous quantities of image data for the species of interest. The conventional approach of sifting through images by eye can be laborious and expensive (although some studies have reduced costs by crowd sourcing; e.g., Hsing et al (2018)). Thus, there is considerable interest in the development of automated methods (Zeppelzauer, 2013).…”