No abstract

A novel method for improving plant disease classification, a challenging and time-consuming process, is proposed. First, using as baseline EfficientNet, a recent and advanced family of architectures having an excellent accuracy/complexity trade-off, we have introduced, devised, and applied refined techniques based on transfer learning, regularization, stratification, weighted metrics, and advanced optimizers in order to achieve improved performance. Then, we go further by introducing adaptive minimal ensembling, which is a unique input to the knowledge base of the proposed solution. This represents a leap forward since it allows improving the accuracy with limited complexity using only two EfficientNet-b0 weak models, performing ensembling on feature vectors by a trainable layer instead of classic aggregation on outputs. To the best of our knowledge, such an approach to ensembling has never been used before in literature. Our method was tested on PlantVillage, a public reference dataset used for benchmarking models' performances for crop disease diagnostic, considering both its original and augmented versions. We noticeably improved the state of the art by achieving 100% accuracy in both the original and augmented datasets. Results were obtained using PyTorch to train, test, and validate the models; reproducibility is granted by providing exhaustive details, including hyperparameters used in the experimentation. A Web interface is also made publicly available to test the proposed methods.

The design of workflow systems originated as an attempt to support coordinated data access. The improvement of interaction technology has created an opportunity for more flexible and interactive activities. Tools for modelling tasks in cooperative applications have started to appear. In this paper, we show how to extend one such tool for supporting workflow control in distributed environments through interactive graphical interfaces. In particular, we show how we have created an environment that exploits a workflow server containing a simulator of cooperative task models. This enables the possibility of allowing users with different roles to access the system through interactive graphical Web interfaces obtained using SVG. Users can access from any location where a Web access is available and obtain information regarding the state of a specific workflow instance, their enabled tasks according to the current state, and the history of tasks accomplished with details regarding their performance. Whenever a task is performed, the user can inform the system regarding this, and the simulator will update the state of the procedure and accordingly enable or disable each user's tasks for all the roles involved.