Crowdsourcing snake identification with online communities of professional herpetologists and avocational snake enthusiasts

Abstract

Species identification can be challenging for biologists, healthcare practitioners and members of the general public. Snakes are no exception, and the potential medical consequences of venomous snake misidentification can be significant. Here, we collected data on identification of 100 snake species by building a week-long online citizen science challenge which attracted more than 1000 participants from around the world. We show that a large community including both professional herpetologists and skilled avocational snake enthusiasts with the potential to quickly (less than 2 min) and accurately (69–90%; see text) identify snakes is active online around the clock, but that only a small fraction of community members are proficient at identifying snakes to the species level, even when provided with the snake's geographical origin. Nevertheless, participants showed great enthusiasm and engagement, and our study provides evidence that innovative citizen science/crowdsourcing approaches can play significant roles in training and building capacity. Although identification by an expert familiar with the local snake fauna will always be the gold standard, we suggest that healthcare workers, clinicians, epidemiologists and other parties interested in snakebite could become more connected to these communities, and that professional herpetologists and skilled avocational snake enthusiasts could organize ways to help connect medical professionals to crowdsourcing platforms. Involving skilled avocational snake enthusiasts in decision making could build the capacity of healthcare workers to identify snakes more quickly, specifically and accurately, and ultimately improve snakebite treatment data and outcomes.

Supervised Learning Computer Vision Benchmark for Snake Species Identification From Photographs: Implications for Herpetology and Global Health

“…When we asked human experts to identify the same images, they performed slightly better overall (species-level F1 0.76 for 45 classes, F1 0.79 for 27 classes; Table 1), with significant variation among species (Durso et al, 2021). Humans correctly labeled the 248 images in TD3 at the species level 53% of the time (N 26,672), although this varied by species from 83% for Agkistrodon contortrix to 30% for Nerodia erythrogaster.…”

“…For the iNaturalist data, only "research grade" observations, which require a community-supported and agreed-upon identification at the species taxonomic rank, were used. Additionally, we selected a subset (N 336) of these images for additional label validation by human experts (Durso et al, 2021) and found that just five (1.5%) were misidentified (these have been corrected on iNaturalist). HerpMapper is used primarily by experienced enthusiasts with a lot of experience in snake identification; we found no misidentifications in the subset (N 200) we examined.…”

“…Subset sample sizes were chosen to represent 0.5% of the total dataset from each source. Human experts (N 250) were recruited via social media, targeting Facebook groups that specialize in snake identification (Durso et al, 2021) as well as by email or private messages over iNaturalist to top identifiers. Because of privacy issues, we were unable to collect demographic information about this community of experts.…”

“…Because of privacy issues, we were unable to collect demographic information about this community of experts. Their accuracy in aggregate was 43% at the species level, 56% at the genus level, and 75% at the family level, with important variation among individuals, snake species, and global region (Durso et al, 2021).…”

“…Many of the images in this dataset were purposefully chosen to be as difficult to identify as possible-e.g., low resolution, out of focus, with the snake filling only a small part of the frame, and/or obscured by vegetation. The identity of species in these images was confirmed by a herpetologist (A. Durso) and they were used in a citizen science challenge where they were presented to participants recruited from online snake identification communities (largely Facebook snake identification groups and iNaturalist) who suggested species identifications, resulting in 68-157 labels of 5-47 classes per image (Durso et al, 2021). We further subdivided TD3 to take all 45 classes into account (TD3a; for fairer comparison with human labeling, because humans were allowed to choose any of the >3,700 snake species classes from a list), and to evaluate only the 27 classes in common (TD3b; for fairer comparison with TD2).…”