CHALLENGE DATA

The dataset is split into two distinct plant species; rice image data and wheat image data. These images were collected at the phenomics facility of Indian Agricultural Research Institute. The dataset comprises of top-view single plant images acquired at different growth stages.

These images have been manually annotated to generate masks for leaf instances both non-occluded and occluded. Specifically, the individual leaf instances and background were labelled with unique grayscale values. In addition, ground truth images with occluded and non-occluded boundaries have also been generated. For each plant image, leaf count has been provided in a CSV file. These annotations have been manually assessed by a plant scientist at Indian Agricultural Research Institute.

To download the dataset first register the challenge and upload the signed EULA form. We will be sending the details to open the dataset.

Training Data 

Test Data