A deep learning approach for Spatio-Temporal forecasting of new cases and new hospital admissions of COVID-19 spread in Reggio Emilia, Northern Italy

Sciannameo, Veronica; Goffi, Alessia; Maffeis, Giuseppe; Gianfreda, Roberta; Pagliari, Daniele Jahier; Filippini, Tommaso; Mancuso, Pamela; Rossi, Paolo Giorgi; Zovo, Leonardo Alberto Dal; Corbari, Angela; Berchialla, Paola

doi:10.1016/j.jbi.2022.104132

Cited by 10 publications

(8 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The strict rules issued by the national government for face mask wearing were the same throughout the country, and were met with high compliance. In contrast, most epidemiologic studies on the role of meteorological factors in COVID-19 spread studied a variety of communities and countries with limited control of covariates, making it difficult to separate the effect of meteorological factors from the effect of confounders and modifying factors, especially mobility and social interaction ( Damette et al, 2021 ; Nottmeyer et al, 2022 ; Sciannameo et al, 2022 ; Wang et al, 2022 ; Zhai et al, 2023 ).…”

Section: Discussionmentioning

confidence: 99%

“…Therefore, there was undoubtedly substantial underascertainment of the incidence and the prevalence of the infection during the first wave, when number of new cases more closely reflected incident COVID-19 cases than new SARS-CoV-2 infections ( Filippini et al, 2021b ). We included the better ascertained outcomes of hospital referrals and deaths not only because they were more reliably measured, but also because meteorological factors might affect not just viral transmission, but also virulence, pathogenicity, and the host response, thus potentially affecting COVID-19 severity and lethality ( Filippini et al, 2021a ; Sciannameo et al, 2022 ). However, the lack of sex-specific data on most confounders precluded the implementation of sex-stratified analysis.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

The influence of meteorological factors on COVID-19 spread in Italy during the first and second wave

Balboni

Filippini

Rothman

et al. 2023

Environmental Research

Self Cite

View full text Add to dashboard Cite

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

The influence of meteorological factors on COVID-19 spread in Italy during the first and second wave

Balboni

Filippini

Rothman

et al. 2023

Environmental Research

Self Cite

View full text Add to dashboard Cite

“…The next multi-frame model explicitly considers the time dependency between frames, replacing the simple feature concatenation with a Long-Short Term Memory (LSTM) cell. Several works have considered CNN-LSTM models to combine spatial and temporal information [48]- [50]. The closest work to ours is [29], which applied a CNN-LSTM for human activity recognition based on a 16x16 IR array.…”

Section: ) Cnn-lstmmentioning

confidence: 99%

Efficient Deep Learning Models for Privacy-Preserving People Counting on Low-Resolution Infrared Arrays

Xie

Daghero

Castellano

et al. 2023

IEEE Internet Things J.

View full text Add to dashboard Cite

Ultra-low-resolution Infrared (IR) array sensors offer a low-cost, energy-efficient, and privacy-preserving solution for people counting, with applications such as occupancy monitoring and visitor flow analysis in private and public spaces. Previous work has shown that Deep Learning (DL) can yield superior performance on this task. However, the literature was missing an extensive comparative analysis of various efficient DL architectures for IR array-based people counting, that considers not only their accuracy, but also the cost of deploying them on memory-and energy-constrained Internet of Things (IoT) edge nodes. Such analysis is key for system designers, since it helps them select the most appropriate DL model given the constraints of their target hardware. In this work, we address this need by comparing 6 different DL architectures on a novel dataset composed of IR images collected from a commercial 8x8 array, which we made openly available. With a wide architectural exploration of each model type, we obtain a rich set of Pareto-optimal solutions, spanning cross-validated balanced accuracy scores in the 55.70-82.70% range. When deployed on a commercial Microcontroller (MCU) by STMicroelectronics, the STM32L4A6ZG, these models occupy 0.41-9.28kB of memory, and require 1.10-7.74ms per inference, while consuming 17.18-120.43 µJ of energy. Our models are significantly more accurate than a previous deterministic method (up to +39.9%), while being up to 3.53x faster and more energy efficient. So, our work serves also as a demonstration that DL can not only achieve higher accuracy, but also higher efficiency compared to classic algorithms for this type of task. Further, our models' accuracy is comparable to state-of-the-art DL solutions on similar resolution sensors, despite a much lower complexity. All our models enable continuous, real-time inference on a MCU-based IoT node, with years of autonomous operation without battery recharging.

show abstract

“…Deep learning techniques may be adopted to create and evaluate arguments, which could have a significant impact on converting digital twin diagnosis and therapy to care delivery. Deep learning methods offer a potential approach to analyze healthcare data sets, assisting in the development of patient-centered health care that can assist individual diagnoses and cure minor medical conditions (Nigo et al 2022, Sciannameo et al 2022. In this light, the present study seeks to investigate the impact of deep learning on a standardised electronic health records dataset for the purpose of detecting kidney disorders.…”

Section: Introductionmentioning

confidence: 99%

Standard-based personalized healthcare delivery for kidney illness using deep learning

Sachdeva

2023

Physiol. Meas.

View full text Add to dashboard Cite

Motivation and Objective: Predictive analytics is one of the activeareas of research in healthcare. It aims to provide better services tothe patient and helps the medical practitioners to know what particulartreatment a patient may need based on their past data. Deep learning isan emerging branch of machine learning in which deep artificial neuralnetworks are used to learn a specific pattern for mapping input to output.It has revolutionized predictive analytics by achieving far better accuracythan conventional learning models. This paper aims to analyze the effectof deep learning on standardized Electronic Health Records dataset bydiagnosing kidney-related diseases.Approach: The current study uses a general modularized deep learningarchitecture called Encoder-Combiner-Decoder (ECD), which offersa robust framework. The model’s performance is enhanced by theavailability of variations and extensions to the basic ECD architecturecorresponding to respective input and output feature types. TheopenEHR Benchmark Dataset (ORBDA) dataset is used to train themodel. It is a real-world dataset that has been provided by the BrazilianPublic Health System through the SUS (DATASUS) DatabaseDepartment of Informatics.Results: In the current research, the model trained using deep learningon the part of this benchmark dataset can help in diagnosing kidneyrelatedillnesses.The evaluation metrics show high precision, recall, andF1 score for kidney-related disease, representing that they can beidentified almost every time.Significance: The model is a novel attempt on analyzing a standardizedhealthcare dataset that can be deployed in medical institutions toevaluate its performance by a medical professional.

show abstract

A deep learning approach for Spatio-Temporal forecasting of new cases and new hospital admissions of COVID-19 spread in Reggio Emilia, Northern Italy

Cited by 10 publications

References 29 publications

The influence of meteorological factors on COVID-19 spread in Italy during the first and second wave

The influence of meteorological factors on COVID-19 spread in Italy during the first and second wave

Efficient Deep Learning Models for Privacy-Preserving People Counting on Low-Resolution Infrared Arrays

Standard-based personalized healthcare delivery for kidney illness using deep learning

Contact Info

Product

Resources

About