The Research

Building a Strong Base in State-of-the-Art Research

Our methods are rooted in SoTA progress: shallow and deep learning, computer vision, AI agency, and multimodal models. We consistently enhance our strategies to guarantee our analyses are founded on the most current global research findings*.

*Visual analysis can offer valuable insights into plant diagnoses, particularly in early detection. However, it should not be viewed as a comprehensive replacement for soil and leaf tissue analysis.

'Using Deep Learning for Image-Based Plant Disease Detection' - 2016

"Crop diseases are a major threat to food security, but their rapid identification remains difficult in many parts of the world due to the lack of the necessary infrastructure. The combination of increasing global smartphone penetration and recent advances in computer vision made possible by deep learning has paved the way for smartphone-assisted disease diagnosis. Using a public dataset of 54,306 images of diseased and healthy plant leaves collected under controlled conditions, we train a deep convolutional neural network to identify 14 crop species and 26 diseases (or absence thereof). The trained model achieves an accuracy of 99.35% on a held-out test set, demonstrating the feasibility of this approach. Overall, the approach of training deep learning models on increasingly large and publicly available image datasets presents a clear path toward smartphone-assisted crop disease diagnosis on a massive global scale."


"Finally, it's worth noting that the approach presented here is not intended to replace existing solutions for disease diagnosis, but rather to supplement them. Laboratory tests are ultimately always more reliable than diagnoses based on visual symptoms alone, and oftentimes early-stage diagnosis via visual inspection alone is challenging. Nevertheless, given the expectation of more than 5 Billion smartphones in the world by 2020—of which almost a Billion in Africa (GSMA Intelligence, 2016)—we do believe that the approach represents a viable additional method to help prevent yield loss. What's more, in the future, image data from a smartphone may be supplemented with location and time information for additional improvements in accuracy. Last but not least, it would be prudent to keep in mind the stunning pace at which mobile technology has developed in the past few years, and will continue to do so. With ever improving number and quality of sensors on mobiles devices, we consider it likely that highly accurate diagnoses via the smartphone are only a question of time."

  • 1Digital Epidemiology Lab, EPFL, Geneva, Switzerland
  • 2School of Life Sciences, EPFL, Lausanne, Switzerland
  • 3School of Computer and Communication Sciences, EPFL, Lausanne, Switzerland
  • 4Department of Entomology, College of Agricultural Sciences, Penn State University, State College, PA, USA
  • 5Department of Biology, Eberly College of Sciences, Penn State University, State College, PA, USA
  • 6Center for Infectious Disease Dynamics, Huck Institutes of Life Sciences, Penn State University, State College, PA, USA


'Plant Disease Identification using Artificial Intelligence: Machine Learning Approach' - 2018

"Computerization in the field of agriculture sees an extraordinary achievement in numerous farming perspectives, including detection of different plant diseases. The focal point of pretty much every nation has moved towards the mechanization of agriculture to achieve exactness and precision and to serve the consistently expanding request of food. Among the significant difficulties in agriculture, plant disease detection is a critical factor influencing the result of cultivating. Quality of vegetables, organic products, vegetables and grains is influenced by plant disease, and hefty misfortune underway and therefore monetary loses are watched, so there is a prerequisite of quick and viable plant disease detection and evaluation strategies. This paper investigates the manners by which machine learning models can be applied to improve the cycle of plant disease detection in beginning phases to improve grain security and manageability of the agro-biological system."


"Applications of machine learning and deep learning in the field of agriculture are picking up energy. Strategies of image preparing are utilized for precise discovery and grouping of harvest disease and the exact location and order of the plan disease’s significant for the productive development of the crop. Several industrially available items are turning out to be well-known step by step to distinguish plant diseases and recognize recuperation arrangements and help farmers in improving their yield profitability and like these benefits."

Jubin Dipakkumar Kothari1
  • 1Department of Information Technology, Campbellsville University, Campbellsville, Kentucky


'How Convolutional Neural Networks Diagnose Plant Disease' - 2019

"Deep learning with convolutional neural networks (CNNs) has achieved great success in the classification of various plant diseases. However, a limited number of studies have elucidated the process of inference, leaving it as an untouchable black box. Revealing the CNN to extract the learned feature as an interpretable form not only ensures its reliability but also enables the validation of the model authenticity and the training dataset by human intervention. In this study, a variety of neuron-wise and layer-wise visualization methods were applied using a CNN, trained with a publicly available plant disease image dataset. We showed that neural networks can capture the colors and textures of lesions specific to respective diseases upon diagnosis, which resembles human decision-making. While several visualization methods were used as they are, others had to be optimized to target a specific layer that fully captures the features to generate consequential outputs. Moreover, by interpreting the generated attention maps, we identified several layers that were not contributing to inference and removed such layers inside the network, decreasing the number of parameters by 75% without affecting the classification accuracy. The results provide an impetus for the CNN black box users in the field of plant science to better understand the diagnosis process and lead to further efficient use of deep learning for plant disease diagnosis."


"In this study, we evaluated an array of visualization methods to interpret the representation of plant diseases that the CNN has diagnosed. The experimental results show that some simple approaches, such as naive visualization of the hidden layer output, are insufficient for plant disease visualization, whereas several state-of-the-art approaches have potential practical applications. Feature visualization and semantic dictionary can be used to extract the visual features that are heavily used to classify a particular disease. To understand what part of the input image is important, the interpretation of attention maps is a favorable choice. However, the behavior of some approaches for generating attention maps was different from what the original study suggested because previous experiments utilized the general object recognition dataset (i.e., ImageNet), which requires the extraction of fine-grained differences, unlike the plant disease diagnosis. Our task is similar to domain-specific fine-grained visual categorization (FGVC) [], which occasionally makes the problem more challenging. This is somewhat related to the datasets of natural images (e.g., iNaturalist dataset []) that contain categories with a similar appearance. It is important to further understand what the deep networks learn for such fine-grained categorization tasks."

Yosuke Toda1,2  Fumio Okura 1,3
  • 1Japan Science and Technology Agency, 4-1-8 Honcho, Kawaguchi, Saitama 332-0012, Japan
  • 2Institute of Transformative Bio-Molecules (WPI-ITbM), Nagoya University, Chikusa, Nagoya 464-8602, Japan
  • 3Department of Intelligent Media, Institute of Scientific and Industrial Research, Osaka University, 8-1 Mihogaoka, Ibaraki, Osaka 567-0047, Japan


'Convolutional Neural Networks for the Automatic Identification of Plant Diseases' - 2019

"Deep learning techniques, and in particular Convolutional Neural Networks (CNNs), have led to significant progress in image processing. Since 2016, many applications for the automatic identification of crop diseases have been developed. These applications could serve as a basis for the development of expertise assistance or automatic screening tools. Such tools could contribute to more sustainable agricultural practices and greater food production security. To assess the potential of these networks for such applications, we survey 19 studies that relied on CNNs to automatically identify crop diseases. We describe their profiles, their main implementation aspects and their performance. Our survey allows us to identify the major issues and shortcomings of works in this research area. We also provide guidelines to improve the use of CNNs in operational contexts as well as some directions for future research."


"In this paper, we identified some of the major issues and shortcomings of works that used CNNs to automatically identify crop diseases. We also provided guidelines and procedures to follow in order to maximize the potential of CNNs deployed in real-world applications. Many already-published solutions based on CNNs are not currently operational for field use mostly due to a lack of conformity to several important concepts of machine learning. This lack of conformity may lead to poor generalization capabilities for unfamiliar data samples and/or imaging conditions, which lowers the practical use of the trained models. Nevertheless, the studied works show the potential of deep learning techniques for crop diseases identification. Their findings are definitely promising for the development of new agricultural tools that could contribute to a more sustainable and secure food production."

  • 1Department of Applied Geomatics, Université de Sherbrooke, Sherbrooke, QC, Canada
  • 2Vision and Imagery Team, Computer Research Institute of Montréal, Montréal, QC, Canada
  • 3Quebec Centre for Biodiversity Science (QCBS), Montreal, QC, Canada


'Review of the State of the Art of Deep Learning for Plant Diseases: A Broad Analysis and Discussion' - 2020

"Deep learning (DL) represents the golden era in the machine learning (ML) domain, and it has gradually become the leading approach in many fields. It is currently playing a vital role in the early detection and classification of plant diseases. The use of ML techniques in this field is viewed as having brought considerable improvement in cultivation productivity sectors, particularly with the recent emergence of DL, which seems to have increased accuracy levels. Recently, many DL architectures have been implemented accompanying visualisation techniques that are essential for determining symptoms and classifying plant diseases. This review investigates and analyses the most recent methods, developed over three years leading up to 2020, for training, augmentation, feature fusion and extraction, recognising and counting crops, and detecting plant diseases, including how these methods can be harnessed to feed deep classifiers and their effects on classifier accuracy."


"This review has discussed and analysed contemporary shallow and deep architectures and their highest achieved accuracy levels for plant disease detection and crop management. The use of realistic datasets, augmentation methods and different pre-training backbone models has also been analysed. Despite the successes that have been achieved in this field, there are still some challenges facing researchers and future orientations to be suggested:

- Mild symptoms of some plant diseases in their early life cycles.
- Some lesion spots have no determined shapes.
- Plant health considerations via monitoring growth and ripeness life cycle of fruits and leaves.
- Automated labelling and auto segmentation of image samples based GANs.
- The usage of hyperspectral data to feed deep classifiers is a recently developed technique that is recommended for the early detection of plant disease life cycles and the healthy leaf life cycle to differentiate it from a diseased leaf.
- Lastly, future work will include several deep learning models for early classification and detection of plant disease due to huge improvements in deep learning models and the availability in plant datasets. Therefore, that will reflect positively on the quality of plants for future generations."

Reem Ibrahim Hasan1,2 Suhaila Mohd Yusuf1  Laith Alzubaidi2,3