19 Jul The capability of artificial intelligence to identify Estonian plants
The aim of this study was to evaluate the accuracy of artificial intelligence based plant image identification applications among the Estonian Flora. Automated identification is relatively new and has not been examined using an extensive dataset. Two applications, Pl@ntNet and Flora Incognita, were used to identify 1501 images from a national database (551 taxa) and 1000 observations from field conditions (281 taxa). The results displayed that the applications were more accurate in the field (success ca 80% both), with the correct species suggested among top five close to 90% of the occasions. Accuracy varied in the database part. Images with flowers were identified better than others, both applications were quicker than manual identification. Flora Incognita was translated into Estonian.
Category: COMPUTING — Country: ESTONIA — Year: 2021
Jaak Pärtel
Video:
Poster:
