Skip to main content
. 2024 Nov 26;6:0272. doi: 10.34133/plantphenomics.0272

Fig. 2.

Fig. 2.

The 2 processes of dataset construction. Step 1: Two experts annotate diagnostic texts for the same image, extracting commonly used disease image description labels and generating expert-diagnosed text templates. Step 2: Employing crowdsourcing techniques, we manually label images and filled in labels into the templates, resulting in the creation of a comprehensive dataset.