Leaf Box is a toolbox written in python developed as a result of the following manuscript

” Fenu, G., & Malloci, F. M. (2021). Using Multioutput Learning to Diagnose Plant Disease and Stress Severity. Complexity, 2021. doi:https://doi.org/10.1155/2021/6663442“

At the state-of-the-art, there are a number of limitations in the disease classification task, including reproducibility of results. As exposed in [1] [2], computational science grapples with reproducibility problems, partly because researchers find it difficult to reproduce algorithm-based results, and partly because of the availability of the datasets themselves. Open implementation increases the likelihood that other researchers can use and develop techniques, and should become common practice. As found in other areas, the release of the free working code helps the research for the intended field.

In agreement with aforementioned observations, the implemented code has been further refined and structured posters was understandable and reusable by others. This commitment has given rise to a toolbox designed for the study of disease prediction based on image analysis through deep learning models, with the hope of providing:

• a support tool in the educational field, to train young students and researchers in the training of deep learning models applied for the intended purposes;
• a modular architecture that speeds up the design of workflows and the integration of new techniques;
• easy-to-use tool that can also be used by users without significant programming experience, allowing the latter to neglect the implementation details and save time, paying more attention to the analysis of the study in question;
• A tool that facilitates the reproducibility of experimental results.

The code is released on Github.

Please remember that our project is in the early stages of development, and we will be committed to integrating new features.

The project can grow together with your contribution. If you are interested, please write to:

francescam [dot] malloci [at] unica [dot] it

References

[1] Does your code stand up to scrutiny?. (2018). Nature, 555(7695), 142. https://doi.org/10.1038/d41586-018-02741-4;

[2] Barnes, N. Publish your computer code: it is good enough. Nature 467, 753 (2010). https://doi.org/10.1038/467753a.