Supervised Learning Computer Vision Benchmark for Snake Species Identification From Photographs: Implications for Herpetology and Global Health
We trained a computer vision algorithm to identify 45 species of snakes from photos and compared its performance to that of humans. Both human and algorithm performance is substantially better than randomly guessing (null probability of guessing correctly given 45 classes = 2.2%). Some species (e.g....
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Frontiers Media S.A.
2021-04-01
|
Series: | Frontiers in Artificial Intelligence |
Subjects: | |
Online Access: | https://www.frontiersin.org/articles/10.3389/frai.2021.582110/full |
id |
doaj-873f9cb0020042629c72fb18a3934b2c |
---|---|
record_format |
Article |
spelling |
doaj-873f9cb0020042629c72fb18a3934b2c2021-04-20T04:47:03ZengFrontiers Media S.A.Frontiers in Artificial Intelligence2624-82122021-04-01410.3389/frai.2021.582110582110Supervised Learning Computer Vision Benchmark for Snake Species Identification From Photographs: Implications for Herpetology and Global HealthAndrew M. Durso0Andrew M. Durso1Gokula Krishnan Moorthy2Sharada P. Mohanty3Isabelle Bolon4Marcel Salathé5Marcel Salathé6Rafael Ruiz de Castañeda7Department of Biological Sciences, Florida Gulf Coast University, Ft. Myers, FL, United StatesInstitute of Global Health, Faculty of Medicine, University of Geneva, Geneva, SwitzerlandEloop Mobility Solutions, Chennai, IndiaAICrowd, Lausanne, SwitzerlandInstitute of Global Health, Faculty of Medicine, University of Geneva, Geneva, SwitzerlandAICrowd, Lausanne, SwitzerlandDigital Epidemiology Laboratory, École Polytechnique Fédérale de Lausanne, Geneva, SwitzerlandInstitute of Global Health, Faculty of Medicine, University of Geneva, Geneva, SwitzerlandWe trained a computer vision algorithm to identify 45 species of snakes from photos and compared its performance to that of humans. Both human and algorithm performance is substantially better than randomly guessing (null probability of guessing correctly given 45 classes = 2.2%). Some species (e.g., Boa constrictor) are routinely identified with ease by both algorithm and humans, whereas other groups of species (e.g., uniform green snakes, blotched brown snakes) are routinely confused. A species complex with largely molecular species delimitation (North American ratsnakes) was the most challenging for computer vision. Humans had an edge at identifying images of poor quality or with visual artifacts. With future improvement, computer vision could play a larger role in snakebite epidemiology, particularly when combined with information about geographic location and input from human experts.https://www.frontiersin.org/articles/10.3389/frai.2021.582110/fullfine-grained image classificationcrowd-sourcingreptilesepidemiologybiodiversity |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Andrew M. Durso Andrew M. Durso Gokula Krishnan Moorthy Sharada P. Mohanty Isabelle Bolon Marcel Salathé Marcel Salathé Rafael Ruiz de Castañeda |
spellingShingle |
Andrew M. Durso Andrew M. Durso Gokula Krishnan Moorthy Sharada P. Mohanty Isabelle Bolon Marcel Salathé Marcel Salathé Rafael Ruiz de Castañeda Supervised Learning Computer Vision Benchmark for Snake Species Identification From Photographs: Implications for Herpetology and Global Health Frontiers in Artificial Intelligence fine-grained image classification crowd-sourcing reptiles epidemiology biodiversity |
author_facet |
Andrew M. Durso Andrew M. Durso Gokula Krishnan Moorthy Sharada P. Mohanty Isabelle Bolon Marcel Salathé Marcel Salathé Rafael Ruiz de Castañeda |
author_sort |
Andrew M. Durso |
title |
Supervised Learning Computer Vision Benchmark for Snake Species Identification From Photographs: Implications for Herpetology and Global Health |
title_short |
Supervised Learning Computer Vision Benchmark for Snake Species Identification From Photographs: Implications for Herpetology and Global Health |
title_full |
Supervised Learning Computer Vision Benchmark for Snake Species Identification From Photographs: Implications for Herpetology and Global Health |
title_fullStr |
Supervised Learning Computer Vision Benchmark for Snake Species Identification From Photographs: Implications for Herpetology and Global Health |
title_full_unstemmed |
Supervised Learning Computer Vision Benchmark for Snake Species Identification From Photographs: Implications for Herpetology and Global Health |
title_sort |
supervised learning computer vision benchmark for snake species identification from photographs: implications for herpetology and global health |
publisher |
Frontiers Media S.A. |
series |
Frontiers in Artificial Intelligence |
issn |
2624-8212 |
publishDate |
2021-04-01 |
description |
We trained a computer vision algorithm to identify 45 species of snakes from photos and compared its performance to that of humans. Both human and algorithm performance is substantially better than randomly guessing (null probability of guessing correctly given 45 classes = 2.2%). Some species (e.g., Boa constrictor) are routinely identified with ease by both algorithm and humans, whereas other groups of species (e.g., uniform green snakes, blotched brown snakes) are routinely confused. A species complex with largely molecular species delimitation (North American ratsnakes) was the most challenging for computer vision. Humans had an edge at identifying images of poor quality or with visual artifacts. With future improvement, computer vision could play a larger role in snakebite epidemiology, particularly when combined with information about geographic location and input from human experts. |
topic |
fine-grained image classification crowd-sourcing reptiles epidemiology biodiversity |
url |
https://www.frontiersin.org/articles/10.3389/frai.2021.582110/full |
work_keys_str_mv |
AT andrewmdurso supervisedlearningcomputervisionbenchmarkforsnakespeciesidentificationfromphotographsimplicationsforherpetologyandglobalhealth AT andrewmdurso supervisedlearningcomputervisionbenchmarkforsnakespeciesidentificationfromphotographsimplicationsforherpetologyandglobalhealth AT gokulakrishnanmoorthy supervisedlearningcomputervisionbenchmarkforsnakespeciesidentificationfromphotographsimplicationsforherpetologyandglobalhealth AT sharadapmohanty supervisedlearningcomputervisionbenchmarkforsnakespeciesidentificationfromphotographsimplicationsforherpetologyandglobalhealth AT isabellebolon supervisedlearningcomputervisionbenchmarkforsnakespeciesidentificationfromphotographsimplicationsforherpetologyandglobalhealth AT marcelsalathe supervisedlearningcomputervisionbenchmarkforsnakespeciesidentificationfromphotographsimplicationsforherpetologyandglobalhealth AT marcelsalathe supervisedlearningcomputervisionbenchmarkforsnakespeciesidentificationfromphotographsimplicationsforherpetologyandglobalhealth AT rafaelruizdecastaneda supervisedlearningcomputervisionbenchmarkforsnakespeciesidentificationfromphotographsimplicationsforherpetologyandglobalhealth |
_version_ |
1721518738128240640 |