The DeepShoot tool was developed to enable fully automated analysis of plant shoot visible light (RGB) images.
Deep learning segmentation models incorporated in DeepShot were separately trained on three different plant types
(Arabidopsis, Maize, Barley) grown in greenhouse facilities and imaged from side and top views.
The main two tasks automatically performed by DeepShoot include (i) binary segmentation of plant shoots
followed by (ii) calculation of of a number of phenotypic traits of segmented plant structures.
For the segmentation task, users should select an appropriate segmentation model (e.g., Arabidopsis/top view or Maize/side view).
Despite the fact that DeepShoot segmentation models were trained on particular types of plants, they can also be applied to
segmentation and phenotyping of other optically similar plants.
 Narisetti N, Henke M, Neumann K, Stolzenburg F, Altmann T and Gladilin E (2022) Deep Learning Based Greenhouse Image Segmentation and Shoot Phenotyping (DeepShoot). Front. Plant Sci. 13:906410. https://doi.org/10.3389/fpls.2022.906410
Leibniz Institute of Plant Genetics and Crop Plant Research (IPK)
06466 Stadt Seeland,
Telefon: +49 (0)39482 50
Telefax: +49 (0)39482 5139