Pasar al contenido principal

Resultados relacionados

Resultados totales (Incluyendo duplicados): 5
Encontrada(s) 1 página(s)

RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Variability of daily feed intake as an indicator of resilience in Pietrain pigs

  • Casto-Rebollo, Cristina|||0000-0003-4646-345X
  • Nuñez-Romano, Pedro
  • Ibáñez-Escriche, Noelia|||0000-0002-6221-3576
  • Gol, S.
  • Reixach, J.
[EN] There is a growing need to produce more resilient livestock that can cope with extreme environments and their associated impacts. Daily feed intake (DFI) is a promising metric for the development of resilience indicators (RIs), as reduced feed consumption is widely recognised as a clinical sign of disease. However, there is no consensus on which DFI-based RIs are the most informative. The objectives of this study were: (i) to evaluate the most common DFI-based RIs and their relationships with feed efficiency traits (FETs), (ii) to investigate the relationship between these indicators and progeny mortality rates on commercial farms, and (iii) to gain insight into their biological mechanisms. A total of 111 121 DFI records of 1 634 healthy purebred Pietrain boars were used, ranging from 47 to 90 days, with an average of 69 DFI records per boar. Three RIs were calculated: residual variance, RMSE and logarithm of the variance of the deviations from an expected pattern (LnVar). A classical animal model was used to estimate the variance components of each RI, and a bivariate model was implemented to estimate the genetic correlation between RIs and the FET. Furthermore, a preliminary mortality study was carried out using data from 6 889 progeny of 55 boars on farms catalogued as virulent for porcine reproductive and respiratory syndrome. Progeny mortality was calculated for each contemporary group (farm and batch) of the offspring, and sires were classified into three levels of resilience. A Bayesian linear model was used to determine whether the differences in progeny mortality rate between the levels of resilience were relevant. On the other hand, a weighted single-step genome-wide association study was performed for each RI and the FET, using a total of 1 216 genotypes. RIs showed moderate heritability = 0.27-0.49) and moderate to strong genetic correlation with FET. Progeny of resilient sires had an 80% probability of at least a 2.5% higher survival rate. Overlapping genomic regions were found for RIs and feed conversion ratio, highlighting DOCK1, SYK and SPTLC1 genes for their potential roles in modulating immune responses and/or metabolism. The LnVar of deviations from the population mean was the most promising indicators (LnVarPop) for disease resilience, as it was the RI that better captured differences in progeny mortality. Furthermore, these results suggest a common biological basis for RIs and FET of interest to breeding programmes. Further studies are needed to validate them. (c) 2025 The Authors. Published by Elsevier B.V. on behalf of The animal Consortium., This work was supported by the Spanish Ministry of Science Innovation in Projects CDTI_ID_20210094, PID2020-114705RB-I00, and TED2021-130759B.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Evaluation of Multi-Object Detection models for automated goat behavior identification in intensive farming facilities

  • Méndez-Reyes, Daniel Alexander|||0000-0001-6414-305X
  • Fajardo-Viloria, Blanca
  • Gil-Vidal, Eliseo
  • Calvet, S.|||0000-0001-6639-9892
  • Costantino, Andrea|||0000-0003-4298-432X
[EN] Computer vision offers significant potential for the continuous, stress-free, and cost-effective monitoring of animal behavior, yet its application in goat farming remains limited. In this study, a Multi-Object Detection (MOD) model was developed to classify goat behavior into four categories: eating, standing, drinking, and lying. Zenithal videos were recorded under both light and dark conditions on an experimental goat farm, resulting in 10740 labelled annotations used to train, validate, and test 13 models leveraging Transformer- and CNN-based pretrained architectures. YOLO-based models (YOLOv8 and YOLOX) achieved the highest overall performances across both large and lightweight versions, demonstrating high detection capability and potential suitability for hardwareconstrained scenarios. YOLOX-based MOD model is identified as the most suitable for goat behavior detection due to its superior classification accuracy, fast inference speed, and fully open-source license, enabling flexible customization, deployment, and reproducibility. Other models, particularly DAB-DETR and H-DINO, underperformed, especially in detecting drinking behavior, which represents the most challenging class due to its visual similarity with standing, class imbalance, and fisheye distortion effects that affect the frame regions where drinkers are located. Mitigation strategies, including focal loss and distortion correction, improved detection accuracy for this class and reduced performance variability. The developed MOD model can be deployed for continuous group-level monitoring of goats, paving the way for scalable and efficient solutions for advanced behavioral analyses. Future works will focus on integrating tracking algorithms for animal-level insights, as well as on evaluating model generalizability across different farming conditions and goat breeds., This work was supported by the "Ayuda a Primeros Proyectos de Investigacion" program [grant number PAID-06-23] , Vice-rectorate for Research of Universitat Politecnica de Valencia (UPV) . Andrea Costantino was supported by the grant JDC2022-049782-I funded by MICIU/AEI/10.13039/501100011033 and by the European Union NextGenerationEU/PRTR. Daniel Alexander Mendez was supported by R+D+i project TED2021-130759B-C31, funded by MCIN/AEI/10.13039/501100011033/and by the "European Union NextGenerationEU/PRTR" and by the grant JDC2023-050534-I funded by MICIU/AEI/10.13039/501100011033 and for FSE+. Eliseo Gil Vidal was supported by PLF4ALL project (CIAICO/2021/212, Generalitat Valenciana).
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Soft Computing Approaches for Predicting Shade-Seeking Behavior in Dairy Cattle Under Heat Stress: A Comparative Study of Random Forests and Neural Networks

  • Sanjuan-Silvestre, Sergi
  • Méndez-Reyes, Daniel Alexander|||0000-0001-6414-305X
  • Arnau, Roger|||0000-0003-2544-8875
  • Calabuig, J. M.|||0000-0001-8398-8664
  • Diaz de Otalora-Aguirre, Xabier|||0000-0003-4841-079X
  • Estellés, F.|||0000-0001-6774-6075
[EN] Heat stress is one of the main welfare and productivity problems faced by dairy cattle in Mediterranean climates. The main objective of this work is to predict heat stress in livestock from shade-seeking behavior captured by computer vision, combined with some climatic features, in a completely non-invasive way. To this end, we evaluate two soft computing algorithms-Random Forests and Neural Networks-clarifying the trade-off between accuracy and interpretability for real-world farm deployment. Data were gathered at a commercial dairy farm in Titaguas (Valencia, Spain) using overhead cameras that counted cows in the shade every 5-10 min during summer 2023. Each record contains the shaded-cow count, ambient temperature, relative humidity, and an exact timestamp. From here, three thermal indices were derived: the current THI, the previous-night mean THI, and the day-time accumulated THI. The resulting dataset covers 75 days and 6907 day-time observations. To evaluate the models' performance a 5-fold cross-validation is also used. The results show that both soft computing models outperform a single Decision Tree baseline. The best Neural Network (3 hidden layers, 16 neurons each, learning rate =10-3) reaches an average RMSE of 14.78, while a Random Forest (10 trees, depth =5) achieves 14.97 and offers the best interpretability. Daily error distributions reveal a median RMSE of 13.84 and confirm that predictions deviate less than one hour from observed shade-seeking peaks. Although the dataset came from a single farm, the results generalized well within the observed range. However, the models could not accurately predict the exact number of cows in the shade. This suggests the influence of other variables not included in the analysis (such as solar radiation or wind data), which opens the door for future research., The research of J.M.C. was funded by the Agencia Estatal de Investigacion, grant number PID2022-138328OB-C21. The research of R.A. was funded by the Universitat Politecnica de Valencia, Programa de Ayudas de Investigacion y Desarrollo (PAID-01-21). The research of D.A.M. was supported by R+D+i project TED2021-130759B-C31, funded by MCIN/AEI/10.13039/501100011033/ and by the "European Union NextGenerationEU/PRTR". This work has received funding from the European Union's Horizon Europe research and innovation program under the grant agreement No. 01059609 (Re-Livestock project).
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Development of goat behaviour prediction with accelerometer data: a machine learning and pre-processing approach

  • Méndez-Reyes, Daniel Alexander|||0000-0001-6414-305X
  • Fajardo-Viloria, Blanca
  • Sanjuan-Silvestre, Sergi
  • Calabuig, J. M.|||0000-0001-8398-8664
  • Arnau, Roger|||0000-0003-2544-8875
  • Calvet, S.|||0000-0001-6639-9892
  • Estellés, F.|||0000-0001-6774-6075
  • Villagra, Arantxa
[EN] The increasing use of accelerometer data for monitoring livestock behaviour in Precision Livestock Farming (PLF) has prompted interest in optimizing machine learning models for real-time applications. This study evaluates the effects of pre-processing factors on predicting goat behaviours using accelerometer data collected in an intensive production environment. A triaxial accelerometer placed on goats' necks recorded movement data, which was synchronized with video-based ethograms for behavioural annotation. Multiple pre-processing techniques, including filtering, windowing, overlapping and sampling frequency with several feature extraction parameters, were assessed to identify optimal combinations for behaviour classification. Various machine learning algorithms, including classification trees, logistic regression, and multilayer perceptron (MLP) models, were applied to predict eating, walking, and inactive behaviours. Results indicate that some of the pre-processing methods applied could induce inflated evaluation metrics and the importance of the selection of train and test sets. Tree-based classifiers and MLPs demonstrate robust performance, achieving average accuracies above 0.9. Battery performance demonstrate that MLP extends the battery life of the accelerometer device by similar to 25 %. These findings highlight the potential of machine learning models in real-time behavioural monitoring to enhance livestock management with goats., This work was supported by R+D+i project TED2021-130759B-C31, funded by MCIN/AEI/10.13039/501100011033/and by the "European Union NextGeneration EU/PRTR".
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Visión artificial para la identificación en tiempo real del comportamiento en cabras

  • Méndez-Reyes, Daniel Alexander|||0000-0001-6414-305X
  • Fajardo-Viloria, Blanca
  • Gil-Vidal, Eliseo
  • Calvet, S.|||0000-0001-6639-9892
  • Costantino, Andrea|||0000-0003-4298-432X
[ES] El comportamiento es clave para evaluar el bienestar y salud de las cabras, puesto que detectar cambios tempranos en sus patrones de actividad permite anticipar posibles problemas sanitarios., El equipo de investigación quiere agradecer a los técnicos José Vicente Martí Vicent y José Luis Palomares Carrasco por su contribución en las actividades de granja. Este trabajo fue financiado con Ayuda a Primeros Proyectos de Investigación (PAID-06-23), Vicerrectorado de Investigación de la Universitat Politècnica de València (UPV) y parte del equipo de investigación benefició de la ayuda del proyecto TED2021-130759B-C31, financiado por MCIN/ AEI/10.13039/501100011033/ y por la "Unión Europea NextGenerationEU/PRTR".