1,721,494 research outputs found
Demographic Fairness in Multimodal Biometrics: A Comparative Analysis on Audio-Visual Speaker Recognition Systems
Full text linkIn urban scenarios, biometric recognition technologies are being increasingly adopted to empower citizens with a secure and usable access to personalized services. Given the challenging environmental scenarios, combining evidence from multiple biometrics at a certain step of the recognition pipeline has been often proved to increase the performance of the biometric-enabled recognition system. Despite the increasing accuracy achieved so far, it still remains under-explored how the adopted biometric fusion policy impacts on the quality of the decisions made by the biometric system, depending on the demographic characteristics of the citizen under consideration. In this paper, we investigate the extent to which state-of-the-art multimodal recognition systems based on facial and vocal biometrics are susceptible to unfairness towards legally-protected groups of individuals, characterized by a common sensitive attribute. Specifically, we present a comparative analysis of the performance across groups for two deep learning architectures tailored for facial and vocal recognition, under seven fusion policies that cover different pipeline steps (feature, model, score and decision). Experiments show that, compared to the unimodal systems alone and the other fusion policies, the multimodal system obtained via a fusion at the model step leads to the highest overall accuracy and the lowest disparity across groups
Forecasting Plant and Crop Disease: An Explorative Study on Current Algorithms
Full text linkEvery year, plant diseases cause a significant loss of valuable food crops around the world. The plant and crop disease management practice implemented in order to mitigate damages have changed considerably. Today, through the application of new information and communication technologies, it is possible to predict the onset or change in the severity of diseases using modern big data analysis techniques. In this paper, we present an analysis and classification of research studies conducted over the past decade that forecast the onset of disease at a pre-symptomatic stage (i.e., symptoms not visible to the naked eye) or at an early stage. We examine the specific approaches and methods adopted, pre-processing techniques and data used, performance metrics, and expected results, highlighting the issues encountered. The results of the study reveal that this practice is still in its infancy and that many barriers need to be overcome
Evaluating Impacts between Laboratory and Field-Collected Datasets for Plant Disease Classification
No full textDeep learning with convolutional neural networks represents the most used approach in recent years in the classification of leaves' diseases. The literature has extensively addressed the problem using laboratory-acquired datasets with a homogeneous background. In this article, we explore the variability factors that influence the classification of plant diseases by analyzing the same plant and disease under different conditions, i.e., in the field and in the laboratory. Two plant species and five biotic stresses are analyzed using different architectures, such as EfficientB0, MobileNetV2, InceptionV2, ResNet50 and VGG16. Experiments show that model performance drops drastically when using representative datasets, and the features learned from the network to determine the class do not always belong to the leaf lesion. In the worst case, the accuracy drops from 92.67% to 54.41%. Our results indicate that while deep learning is an effective technique, there are some technical issues to consider when applying it to more representative datasets collected in the field
DiaMOS Plant: a dataset for diagnosis and monitoring plant disease
No full textThe classification and recognition of foliar diseases is an increasingly developing field of research, where the concepts of machine and deep learning are used to support agricultural stakeholders. Datasets are the fuel for the development of these technologies. In this paper, we release and make publicly available the field dataset collected to diagnose and monitor plant symptoms, called DiaMOS Plant, consisting of 3505 images of pear fruit and leaves affected by four diseases. In addition, we perform a comparative analysis of existing literature datasets designed for the classification and recognition of leaf diseases, highlighting the main features that maximize the value and information content of the collected data. This study provides guidelines that will be useful to the research community in the context of the selection and construction of datasets
Using Multioutput Learning to Diagnose Plant Disease and Stress Severity
Full text linkEarly diagnosis of leaf diseases is a fundamental tool in precision agriculture, thanks to its high correlation with food safety and environmental sustainability. It is proven that plant diseases are responsible for serious economic losses every year. The aim of this work is to study an efficient network capable of assisting farmers in recognizing pear leaf symptoms and providing targeted information for rational use of pesticides. The proposed model consists of a multioutput system based on convolutional neural networks. The deep learning approach considers five pretrained CNN architectures, namely, VGG-16, VGG-19, ResNet50, InceptionV3, MobileNetV2, and EfficientNetB0, as feature extractors to classify three diseases and six severity levels. Computational experiments are conducted to evaluate the model on the DiaMOS Plant dataset, a self-collected dataset in the field. The results obtained confirm the robustness of the proposed model in automatically extracting the discriminating features of diseased leaves by adopting the multitasking learning paradigm
DSS LANDS: A Decision Support System for Agriculture in Sardinia
No full textRecently, the use of DSSs application has been strongly increasing in the agricultural sector due to continuous climate change and the need to conduct more productive and sustainable agriculture. In this paper, we describe the prototype agricultural DSS LANDS developed for monitoring the main crop production in Sardinia. The DSS collects, organizes, integrates, and analyzes several types of data with different mathematical models. In particular, a case study on forecasting potato late blight is presented. We employed the Negative Prognosis model and the Fry model to forecast the period in which it is opportune to carry out fungicide treatments useful against the appearance of the pathogen. The experiments allowed us to outline the best criteria for local conditions, and the evaluation showed the effectiveness of the approach in a concrete case study
Classification of Pear Leaf Diseases Based on Ensemble Convolutional Neural Networks
Full text linkOver the last few years, the impact of climate change has increased rapidly. It is influencing all steps of plant production and forcing farmers to change and adapt their crop management practices using new technologies based on data analytics. This study aims to classify plant diseases based on images collected directly in the field using deep learning. To this end, an ensemble learning paradigm is investigated to build a robust network in order to predict four different pear leaf diseases. Several convolutional neural network architectures, named EfficientNetB0, InceptionV3, MobileNetV2 and VGG19, were compared and ensembled to improve the predictive performance by adopting the bagging strategy and weighted averaging. Quantitative experiments were conducted to evaluate the model on the DiaMOS Plant dataset, a self-collected dataset in the field. Data augmentation was adopted to improve the generalization of the model. The results, evaluated with a range of metrics, including accuracy, recall, precison and f1-score, showed that the proposed ensemble convolutional neural network outperformed the single convolutional neural network in classifying diseases in real field-condition with variation in brightness, disease similarity, complex background, and multiple leaves
Artificial intelligence technique in crop disease forecasting: a case study on potato late blight prediction
No full textCrop diseases are strongly affected by weather and environmental factors. Weather fluctuations represent the main factors that lead to potential economic losses. The integration of forecasting models based on weather data can provide a framework for agricultural decision-making able to suggest key information for overcoming these problems. In the present work, we propose a new artificial intelligence approach to forecast potato late blight disease in the Sardinia region and a novel technique to express a crop disease risk. The experiments conducted are based on historical weather data as temperature, humidity, rainfall, speed wind, and solar radiation collected from several locations over 4 year (2016–2019). The tests were aimed to determine the usefulness of the support vector machine classifier to identify crop–weather–disease relations for potato crops and the relative possible outbreak. The results obtained show that temperature, humidity, and speed wind play a key role in the prediction
An application of machine learning technique in forecasting crop disease
No full textIn the recent years, Big Data Analytics and Machine Learning techniques are playing an increasingly key role in the agriculture sector in order to tackle the increasing challenges due to the climate changes which are causing serious damage production. The analysis of environmental, climatic and cultural factors allows to establish the irrigation and nutritional needs of crops, forecast crop disease, improve crop yield, as well as improve the quantity and the quality of agricultural output while using less input. Potato late blight is considered one of the most devasting disease world over, including Sardinia. Unexpected epidemics can result in significant economic and yield losses. In this paper, we describe the test conducted using the DSS LANDS in order to predict potato late blight disease in Sardinia. The object of the study was to investigate if regional weather variables could be used to predict potato late blight risk in southern Sardinia using a Machine Learning approach. The disease severity is predicted using Feed-forward Neural Network and Support Vector Machine Classification based on meteorological parameters provided by ARPAS weather stations. The prediction accuracy for ANN was 96% and for SVM Classification was 98%
Lands DSS: A Decision Support System for Forecasting Crop Disease in Southern Sardinia
No full textDecision support systems (DSSs) are used in precision farming to address climate and environmental changes due to human action. However, increments in the amount of data produced continuously by the latest sensor and satellite technologies have recently incentivized the integration of artificial intelligence (AI). A review of research dedicated to the application of DSSs and AI in forecasting crop disease is proposed. In this paper, the authors describe the DSS LANDS developed for monitoring the main crop productions in Sardinia and the case study conducted to forecast potato late blight. A feed-forward neural network was implemented to investigate if weather data provided by regional stations could be used to predict a disease risk index using an AI technique. The test performed by stratified k-fold cross validation achieved an accuracy of 96%
- …
