Datasets Overview

The Figure symbolically depicts the data available together with ground truth segmentations and further annotations and metadata.

The data have been acquired in two of laboratories with two different setups.  Overall we should images of plants taken from a camera on top. The images show plants of tobacco or Arabidopsis imaged at once or many in the scene.  

The data have been carefully annotated and verified. They can be used for a series of tasks as outlined in the introduction. For examples of images/tasks please click here.

Further details are provided in the main paper:

M. Minervini, A. Fischbach, H.Scharr, and S.A. Tsaftaris. Finely-grained annotated datasets for image-based plant phenotyping. Pattern Recognition Letters, pages 1-10, 2015, doi:10.1016/j.patrec.2015.10.013 [PDF] [BibTex]

If you are interested on the exact experimental protocol additional details are given at:

H. Scharr, M. Minervini, A. Fischbach, S.A. Tsaftaris. Annotated Image Datasets of Rosette Plants. Technical Report No. FZJ-2014-03837, Forschungszentrum Jülich, 2014 [PDF][BibTex]

Additional useful papers are in the Citation page.

Please note that parts of the data have already been used to support two challenges (on leaf segmentation and counting) at the Computer Vision Problems in Plant Phenotyping workshops that occurred in conjunction with ECCV 2014 (Zurich), BMVC 2015 (Swansea), and ICCV2017 (Venice).


Sponsored by




Phidas logo

Some example images from data set