Dataset Information

Computer Vision Applied to Detect Lethargy through Animal Motion Monitoring: A Trial on African Swine Fever in Wild Boar.

ABSTRACT: Early detection of infectious diseases is the most cost-effective strategy in disease surveillance for reducing the risk of outbreaks. Latest deep learning and computer vision improvements are powerful tools that potentially open up a new field of research in epidemiology and disease control. These techniques were used here to develop an algorithm aimed to track and compute animal motion in real time. This algorithm was used in experimental trials in order to assess African swine fever (ASF) infection course in Eurasian wild boar. Overall, the outcomes showed negative correlation between motion reduction and fever caused by ASF infection. In addition, infected animals computed significant lower movements compared to uninfected animals. The obtained results suggest that a motion monitoring system based on artificial vision may be used in indoors to trigger suspicions of fever. It would help farmers and animal health services to detect early clinical signs compatible with infectious diseases. This technology shows a promising non-intrusive, economic and real time solution in the livestock industry with especial interest in ASF, considering the current concern in the world pig industry.

SUBMITTER: Fernandez-Carrion E

PROVIDER: S-EPMC7760671 | biostudies-literature | 2020 Nov

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Computer Vision Applied to Detect Lethargy through Animal Motion Monitoring: A Trial on African Swine Fever in Wild Boar.

Fernández-Carrión Eduardo E Barasona Jose Ángel JÁ Sánchez Ángel Á Jurado Cristina C Cadenas-Fernández Estefanía E Sánchez-Vizcaíno José Manuel JM

Animals : an open access journal from MDPI 20201129 12

Early detection of infectious diseases is the most cost-effective strategy in disease surveillance for reducing the risk of outbreaks. Latest deep learning and computer vision improvements are powerful tools that potentially open up a new field of research in epidemiology and disease control. These techniques were used here to develop an algorithm aimed to track and compute animal motion in real time. This algorithm was used in experimental trials in order to assess African swine fever (ASF) inf ...[more]

PMID: 33260362

Similar Datasets

Project description:Two groups with three wild boars each were used: Group A (animals 1 to 3) served as the control, and Group B (animals 4 to 6) was postnatally persistently infected with the Cat01 strain of CSFV (primary virus). The animals, six weeks old and clinically healthy, were inoculated with the virulent strain Margarita (secondary virus). For exclusive detection of the Margarita strain, a specific qRT-PCR assay was designed, which proved not to have cross-reactivity with the Cat01 strain. The wild boars persistently infected with CSFV were protected from superinfection by the virulent CSFV Margarita strain, as evidenced by the absence of clinical signs and the absence of Margarita RNA detection in serum, swabs and tissue samples. Additionally, in PBMCs, a well-known target for CSFV viral replication, only the primary infecting virus RNA (Cat01 strain) could be detected, even after the isolation in ST cells, demonstrating SIE at the tissue level in vivo. Furthermore, the data analysis of the Margarita qRT-PCR, by means of calculated ΔCt values, supported that PBMCs from persistently infected animals were substantially protected from superinfection after in vitro inoculation with the Margarita virus strain, while this virus was able to infect naive PBMCs efficiently. In parallel, IFN-α values were undetectable in the sera from animals in Group B after inoculation with the CSFV Margarita strain. Furthermore, these animals were unable to elicit adaptive humoral (no E2-specific or neutralising antibodies) or cellular immune responses (in terms of IFN-γ-producing cells) after inoculation with the second virus. Finally, a sequence analysis could not detect CSFV Margarita RNA in the samples tested from Group B. Our results suggested that the SIE phenomenon might be involved in the evolution and phylogeny of the virus, as well as in CSFV control by vaccination. To the best of our knowledge, this study was one of the first showing efficient suppression of superinfection in animals, especially in the absence of IFN-α, which might be associated with the lack of innate immune mechanisms.

Project description:The first cases of African swine fever (ASF) were detected in the Lithuanian wild boar population in 2014. Since then, the disease spread slowly through the whole country, affecting both, wild boar and domestic pigs. In the other Baltic states, which both are also affected by ASF since 2014, the recent course of ASF prevalence suggests that the countries might be well under way of disease elimination. In contrast, in Lithuania the epidemic seems to be still in full progress. In the present study, we aimed to extend a previous prevalence study in Lithuania. Looking at ASF virus (ASFV) and seroprevalence estimates of wild boar in all months of 2018 and in all affected municipalities in Lithuania, the course of ASF was evaluated on a temporal and spatial scale. A non-spatial beta-binomial model was used to correct for under- or overestimation of the average prevalence estimates. Within 2018 no big differences between the prevalence estimates were seen over time. Despite of the lower sample size, highest ASFV prevalence estimates were found in dead wild boar, suggesting higher detection rates through passive surveillance than through active surveillance. Accordingly, with the maximum prevalence of 87.5% in May 2018, the ASFV prevalence estimates were very high in wild boar found dead. The number of samples originating from hunted animals (active surveillance) predominated clearly. However, the ASFV prevalence in those animals was lower with a maximum value of 2.1%, emphasizing the high value of passive surveillance. A slight increase of the seroprevalence in hunted wild boar could be seen over time. In the center of Lithuania, a cluster of municipalities with high ASFV and seroprevalence estimates was found. The results of the study indicate that ASFV is still circulating within the Lithuanian wild boar population, constituting a permanent risk of disease transmission into domestic pig holdings. However, additional, more recent data analyses are necessary to re-evaluate the course of ASF in Lithuania and thus, to be able to make a statement about the stage of the ASF epidemic in the country. This is of huge importance for Lithuania for evaluating control measures and their efficacy, but also for neighbouring countries to assess the risk of disease spread from Lithuania.

Dataset Information

Computer Vision Applied to Detect Lethargy through Animal Motion Monitoring: A Trial on African Swine Fever in Wild Boar.

Publications

Computer Vision Applied to Detect Lethargy through Animal Motion Monitoring: A Trial on African Swine Fever in Wild Boar.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets