Journal Article

Swipe Right: a Comparison of Accuracy of Plant Identification Apps for Toxic Plants

Otter Jenna, Tomaszewski, Christian A, Mayer, Stephanie

2021

Overview

IntroductionPlant identification applications for use on smartphones have been increasing in availability, accuracy, and utilization. We aimed to perform an introductory study to determine if a plant identification application (ID app) used on a smartphone could identify toxic plants, and to compare apps to determine which is most reliable.MethodsWe compared three popular iPhone plant ID apps, PictureThis (PT), PlantSnap (PS), and Pl@ntNet (PN), used to identify 17 commonly encountered toxic plants. Apps were used to photograph the entire plant, leaves, and flowers of ≥ 10 different plants for each species. Two toxicologists performed plant identification with confirmation of identification performed by a botanist, and inter-researcher agreement was confirmed. For each plant species, scores for accuracy of app identification of leaves, flowers, and whole plant were combined to create an overall composite score used to compare accuracy of each app (95% C.I.).ResultsPictureThis had the best performance with 10/17 (59% [36 to 78]) plant species identified 100% correctly, as opposed to 8/17 (47% [26 to 69]) for Pl@ntNet and 1/17 for PlantSnap (5.8% [1.1 to 27]).ConclusionA plant identification app may be a useful tool to assist healthcare providers and the public in identifying toxic plants.

• Publisher: Springer Nature B.V
• Subject: Accuracy / Applications programs / Flowers & plants / Identification / Leaves / Plant species / Smartphones / Species / Toxicology