AI-Assisted Early Detection of Disease Using Bioacoustic and Thermal Imaging Signatures in Wildlife and Livestock
DOI:
https://doi.org/10.62497/ijrab.196Keywords:
AI, bioacoustic monitoring, thermal imaging, early disease detection, livestock, wildlife, non-invasive diagnosticsAbstract
Background: Early detection of disease in wildlife and livestock is critical for animal welfare, biodiversity conservation, and agricultural productivity. Conventional diagnostics are often invasive, time-consuming, or impractical in field conditions, highlighting the need for non-invasive, rapid surveillance tools.
Objective: To develop and evaluate an AI-based system integrating bioacoustic and thermal imaging signatures for the early detection of disease in wildlife and livestock.
Methodology: A prospective observational study was conducted over two years, involving 180 animals (68 goats, 54 sheep, 38 cattle, and 20 wildlife species). Paired bioacoustic and thermal imaging data were collected under standardized conditions. Acoustic features—including fundamental frequency, spectral centroid, amplitude variability, and harmonic-to-noise ratio—alongside thermal parameters such as maximum surface temperature, temperature gradient, thermal asymmetry, and hotspot area, were extracted. Machine learning models were trained to classify animals as healthy or diseased, and performance metrics including accuracy, sensitivity, specificity, F1 score, and area under the ROC curve (AUC) were calculated.
Results: Diseased animals showed significantly higher fundamental frequency (322.78 ± 44.51 Hz vs. 285.42 ± 38.16 Hz), spectral centroid (1788.34 ± 238.56 Hz vs. 1530.67 ± 210.45 Hz), amplitude variability (19.87 ± 4.11% vs. 14.22 ± 3.56%), and lower harmonic-to-noise ratio (3.11 ± 0.92 dB vs. 4.82 ± 1.14 dB). Thermal imaging revealed elevated maximum surface temperature (38.21 ± 1.14°C vs. 35.48 ± 0.92°C), higher temperature gradients (3.12 ± 0.51°C vs. 1.84 ± 0.42°C), greater thermal asymmetry (11.78 ± 2.05% vs. 6.21 ± 1.34%), and larger hotspot areas (7.92 ± 1.88 cm² vs. 4.18 ± 1.21 cm²). The multimodal AI model achieved the highest diagnostic performance (accuracy 93.44%, sensitivity 91.78%, specificity 94.92%, F1 score 0.92, AUC 0.97).
Conclusion: Integration of bioacoustic and thermal imaging with AI provides a highly accurate, non-invasive tool for early disease detection in livestock and wildlife, supporting proactive health monitoring and timely interventions.
References
Lefrançois T, Pineau T. Public health and livestock: emerging diseases in food animals. Animal Frontiers. 2014 Jan 1;4(1):4-6. https://doi.org/10.2527/af.2014-0001.
Scott ME. The impact of infection and disease on animal populations: implications for conservation biology. Conservation biology. 1988 Mar;2(1):40-56. https://doi.org/10.1111/j.1523-1739.1988.tb00334.x
Bartlow AW, Machalaba C, Karesh WB, Fair JM. Biodiversity and global health: intersection of health, security, and the environment. Health security. 2021 Apr 1;19(2):214-22. https://doi.org/10.1089/hs.2020.0112
Bora M, Mathew DD, Das H, Bora DP, Barman NN. Point of care diagnostics and non-invasive sampling strategy: a review on major advances in veterinary diagnostics. Acta Veterinaria Brno. 2022 Feb 3;91(1):17-34. https://doi.org/10.2754/avb202291010017.
Bradley RD, Roberts EK, Worsham AE, Ashton MN, Wright EA, Saleh N, Hardy DM. A Non-invasive, Biomarker Assay for Detecting Chronic Wasting Disease Pathology in White-tailed Deer. bioRxiv. 2022 Jul 21:2022-07. https://doi.org/10.1101/2022.07.20.500695.
Neethirajan S. Transforming the adaptation physiology of farm animals through sensors. Animals. 2020 Aug 26;10(9):1512. https://doi.org/10.3390/ani10091512.
Johnson E, Campos-Cerqueira M, Jumail A, Yusni AS, Salgado-Lynn M, Fornace K. Applications and advances in acoustic monitoring for infectious disease epidemiology. Trends in Parasitology. 2023 May 1;39(5):386-99. https://doi.org/10.1002/ieam.4324
Mcloughlin MP, Stewart R, McElligott AG. Automated bioacoustics: methods in ecology and conservation and their potential for animal welfare monitoring. Journal of the Royal Society Interface. 2019 Jun 28;16(155):20190225. https://doi.org/10.1098/rsif.2019.0225
Yang B, Li X, Hou Y, Meier A, Cheng X, Choi JH, Wang F, Wang H, Wagner A, Yan D, Li A. Non-invasive (non-contact) measurements of human thermal physiology signals and thermal comfort/discomfort poses-a review. Energy and Buildings. 2020 Oct 1;224:110261. https://doi.org/10.1016/j.enbuild.2020.110261.
Xu, Y., Zhang, X., Li, H., Zheng, H., Zhang, J., Olsen, M. S., ... & Qian, Q. (2022). Smart breeding driven by big data, artificial intelligence, and integrated genomic-enviromic prediction. Molecular plant, 15(11), 1664-1695. https://doi.org/10.1016/j.molp.2022.09.001
An Q, Rahman S, Zhou J, Kang JJ. A comprehensive review on machine learning in healthcare industry: classification, restrictions, opportunities and challenges. Sensors. 2023 Apr 22;23(9):4178. https://doi.org/10.3390/s23094178.
MacAulay S, Ellison AR, Kille P, Cable J. Moving towards improved surveillance and earlier diagnosis of aquatic pathogens: From traditional methods to emerging technologies. Reviews in Aquaculture. 2022 Sep;14(4):1813-29. https://doi.org/10.1111/raq.12674
Hoffmann G, Schmidt M, Ammon C, Rose-Meierhöfer S, Burfeind O, Heuwieser W, Berg W. Monitoring the body temperature of cows and calves using video recordings from an infrared thermography camera. Veterinary research communications. 2013 Jun;37(2):91-9. https://doi.org/10.1007/s11259-012-9549-3
Sejian V, Shashank CG, Silpa MV, Madhusoodan AP, Devaraj C, Koenig S. Non-invasive methods of quantifying heat stress response in farm animals with special reference to dairy cattle. Atmosphere. 2022 Oct 9;13(10):1642. https://doi.org/10.3390/atmos13101642.
Astill J, Dara RA, Fraser ED, Sharif S. Detecting and predicting emerging disease in poultry with the implementation of new technologies and big data: A focus on avian influenza virus. Frontiers in veterinary science. 2018 Oct 30;5:263. https://doi.org/10.3389/fvets.2018.00263.
Cuan K, Zhang T, Li Z, Huang J, Ding Y, Fang C. Automatic Newcastle disease detection using sound technology and deep learning method. Computers and Electronics in Agriculture. 2022 Mar 1;194:106740. https://doi.org/10.1016/j.compag.2022.106740.
Cuan K, Zhang T, Li Z, Huang J, Ding Y, Fang C. Automatic Newcastle disease detection using sound technology and deep learning method. Computers and Electronics in Agriculture. 2022 Mar 1;194:106740. https://doi.org/10.3390/s22030705.
Published
Issue
Section
License
Copyright (c) 2024 Nibsha Mahboob, Muhammad Abdullah Khan, Abdus Sami, Muhammad Arslan Latif, Muhammad Usman, Muhammad Kamran Tahir (Author)
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
