The Agricultural AI Team announces the development of a classification system focused on Matthiola incana(M. incana). This ornamental flower variety is utilized in various contexts, such as theme parks and ceremonial occasions. A significant challenge in cultivating M. incana is the process known as double flower identification. This involves the removal of single-flower specimens to cultivate only double-flower individuals, which have greater aesthetic appeal and commercial value. This selection process optimizes resource usage, including space and labor. However, accurately distinguishing between single and double flower specimens requires a level of expertise that can be difficult for less experienced workers.

In collaboration with horticultural experts and students from Tokyo Metropolitan Mizuho Nogei High School, the Nakajo Laboratory has developed the MiLIC (Matthiola incana Leaf Image Classification) dataset. This dataset is designed to support the creation of image classification models for double flower identification through the analysis of images of single and double cotyledons. We have published this dataset on Hugging Face, a platform for machine learning models and datasets. To access the MiLIC dataset, please complete the following Google Form. Upon completion, you will receive a link to download the dataset.

Google Form Link: https://forms.gle/dKmRXPUJsWgSAZq5A

Additionally, a research paper detailing the development of a double flower identification model using the MiLIC dataset is forthcoming. The dataset is available under the CC BY 4.0 license, allowing for use in accordance with its terms.

---