Escrever um comentário

This research article titled “Enhanced Convolutional Networks for Accurate LeafBased Plant Disease Classification” is a well written manuscript that proposes a novel machine learning model. The title, along with the abstract and the keywords are explanatory and without unnecessary information. The introduction and related work sections, are sufficiently documented with introductory information, providing the basic principles and background information and similar past studies and gaps of technologies that the authors attempt to address, respectively. The Enhanced Convolutional Neural Network (ECNN) model, which comprises the main topic of the study is introduced and its structure is clearly defined.

For the remaining parts of the study, the ECNN is compared with other relevant models, presenting significant positive results and thoroughly tested, not only in the popular PlantVillage dataset, but also in field-collected images and synthetic augmentation samples, that have been generated via data augmentation techniques. One drawback is that there is no available information for the hardware used for all the computational processes like model training, validation and augmentations which is very important for this study’s materials and methods justification, and also for repetition purposes. Another possible drawback could be that it is not provided citation of the figures in the text, in order to assist the reader when goes through the study.

Moreover, I would like to propose an enrichment of the figure 1, in terms of graphical design, just for visualization purposes and user experience. Last minor comment is that there is a typo error in a table name (it is written Table X) in the Performance Analysis sub-section in Results and Discussion section. Finally, the conclusion and future work part has some strong points that enhance the study’s credibility and reflect its current limitations, like its light weight in order to be easily deployed in a mobile app and employing other types of image sensors, respectively.

Competing interests

The author declares that they have no competing interests.

Use of Artificial Intelligence (AI)

The author declares that they did not use generative AI to come up with new ideas for their review.