Plant id from photo desktop free#

2018 Wäldchen and Mäder 2018 Wäldchen et al. 2016) and this approach is widely used in the current apps as outlined in many recent papers and reviews that outline progress with development of image recognition for plant identification ( Mata-Montero and Carranza-Rojas 2016 Cappellato et al. Other work has confirmed that deep neural networks can be extraordinarily powerful at detailed recognition even using noisy data ( Krause et al. 2012), for example, was an early automated app for species identification using a large data base of leaf images and provides the basis for PlantSnap evaluated here. Increasing the size of the database of images which can be used for identification has been one key to the rapid recent advances in the field. 2017) all aiming to compare the performance of ‘experts’ with that of the best deep learning algorithms ( Bonnet et al. 2017), with eight different research groups contributing their models to LifeCLEF2017 ( Goëau et al. The sub-project PlantCLEF focussed on plant identification ( Goëau et al. Much effort has been centred on the Cross Language Evaluation Forum (CLEF) initiative ( ), which since 2013 included the LifeCLEF challenge to develop automated identification systems. 2013 Shayan and Sajeevan 2013 Hsaio et al. 2006 Valliammal and Geethalakshmi 2011 Hati and Sajeevan 2013 Satti et al. Many of the earlier studies of automated identification focussed on images of single organs, especially leaves, imaged on plain backgrounds ( Wang et al.

In spite of these challenges, there has been much progress on the development of image recognition/AI approaches to plant identification over the past 20 years with many hundred academic papers addressing various aspects of this problem (see e.g. The third difficulty, particularly relevant for conservation studies, is posed by the rarity of some species which means that they may be absent (or poorly represented) in any reference set of images used for identification. Any viable system needs to address both these problems. A second challenge is that the characters distinguishing between species are often cryptic either being microscopic or requiring very specific views that may not be present in the images available. One is posed by the variability in view and quality of images taken by users in the field. There are at least three major challenges facing all automated systems. There has been rapid recent development of smartphone apps to aid plant identification in the field, ranging from the use of those based on automated image recognition or AI (the subject of this paper), to those that require the user to use traditional dichotomous keys or multi-access keys and those that only provide a selection of images without a clear system for species identification. Species identification is also of interest to a more general audience including nature lovers, hikers and eco-tourists. 2019), thus mitigating this ‘taxonomic gap’ ( Joly et al. It has therefore been suggested that artificial intelligence (AI) might replace such human expertise ( Wäldchen et al. Research into biodiversity and conservation requires species identification, but the availability of botanists with good plant identification skills is declining. Potential uses of automated image recognition plant identification apps are discussed and recommendations made for their future use.Īpps, artificial intelligence, image recognition, plant identification, smartphones Introduction Nevertheless, the better-performing apps should be of great value to beginners and amateurs and may usefully stimulate interest in plant identification and nature. Although the accuracy is good for some of the top-rated apps, for any quantitative biodiversity study or for ecological surveys, there remains a need for validation by experts or against conventional floras. There were large differences in performance with the best apps identifying >50 % of samples tested to genus or better.

Plant id from photo desktop free#

Nine free apps or websites for automated plant identification and suitable for use on mobile phones or tablet computers in the field were tested on a disparate set of 38 images of plants or parts of plants chosen from the higher plant flora of Britain and Ireland. There has been a recent explosion in development of image recognition technology and its application to automated plant identification, so it is timely to consider its potential for field botany.