Title of the dataset: Rob2Pheno Tomato Image Dataset
...

Creators: Manya Afonso, Hubert Fonteijn, Felipe S Fiorentin, Dick Lensink, Marcel Mooij, Nanne Faber, Gerrit Polder, and Ron Wehrens
...

Related publication: Afonso, Manya, et al. "Tomato Fruit Detection and Counting in Greenhouses Using Deep Learning." Frontiers in plant science 11 (2020): 1759. https://doi.org/10.3389/fpls.2020.571299
...

Description: Dataset used to train an object instance detector, to detect tomato fruits in a greenhouse setting.
...

Keywords: tomato, greenhouse, computer vision in agriculturre, deep learning, phenotyping
...

This dataset contains the following files: RGB.tar.gz, Depth.tar.gz, train_1class.JSON, val_1class.JSON, train_2class.JSON, val_2class.JSON
...

Explanation of variables: The JSON files contain the object instance annotations for tomato fruits, in the MS COCO format. The 1 class files assume that all tomato fruits belong to 1 class called 'fruit', while the 2 class ones contain 2 ripeness classes, red (ripe) and green (unripe) fruit.
...

Methods, materials and software: The dataset is to be used with instance object detectors such as MaskRCNN or YOLACT.
...

This dataset is published under the CC BY (Attribution) license.
This license allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator.
