The use of behavioural biometrics, 3D kinematics and infrared thermography to assess affective and biological states in livestock

使用行为生物识别、3D 运动学和红外热成像技术评估牲畜的情感和生物状态

• 批准号: RGPIN-2021-02890
• 负责人: Bench, Clover
• 金额: $ 2.04万
• 依托单位: University of Alberta
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2022
• 资助国家: 加拿大
• 起止时间: 2022-01-01 至 2023-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=743330
• 关键词: use; behavioural; biometrics; 3D; kinematics

As a result of the move away from the Five Freedoms paradigm of welfare assessment, towards a Five Domains model, the field of ethology is challenged to develop valid, reliable, and feasible ways to evaluate a variety of affects experienced by animals. Livestock species are at the forefront of this development as an integral part of the global food system. Since 2016, studies have identified changes in facial action units through the use of grimace scales in animals to measure negative affects, however there are limitations to the use of grimace scales in livestock. For example, individual expression and magnitude of pain can vary between animals, which grimace scales have difficulty differentiating. For these and other reasons, grimace scales for pain assessment in some livestock species, such as pigs, have not been fully validated. My research program is unique as the only program investigating micro-behaviour biometrics in livestock as part of automated technology platforms for use in precision agriculture. The overall aim of my research program has been to better understand how qualitative and quantitative differences in livestock micro-behaviour biometrics can be incorporated into automated systems and combined with machine learning for a variety of research and on-farm applications. In addition to the use of automated infrared thermography platforms to detect micro-behaviour biometrics alongside thermal profiles as part of early disease detection, my research has also used 3D kinematic motion capture technology to better understand the biomechanics of micro-behaviours. For example, the kinematics of micro-behaviours exhibited during estrus, which confirmed micro-movements are better at detecting changes in biological functioning. However, an area that has not been explored is the use of micro-behaviour biometrics to assess affective states and facial expressions in livestock. The focus of the proposed research is to answer an emerging question in animal welfare: Are micro-facial expressions and other micro-behaviour biometrics an epiphenomenon for pain or other affective states in livestock? If so, what are the underlying physiological mechanisms? The short-term objectives of this research are: 1) to characterize the micro-facial expressions associated with positive and negative affective states, and 2) to ascertain the physiological mechanisms (pathway) underlying micro-facial expressions as affective biometrics using 3D kinematic analysis, thermal imaging, and genomic analysis. The proposed research will clarify the role physiology plays in driving micro-facial expressions in response to different types of stimuli experienced by pigs and not only strives to find a solution to current problems with the use of pain grimace scales, but also suggests it can be done using automation integrating machine learning as the Canadian agriculture sector moves towards greater inclusion of precision systems.

由于从福利评估的五个自由范式转向五个领域模型，动物行为学领域面临着开发有效，可靠和可行的方法来评估动物所经历的各种影响的挑战。畜牧物种作为全球粮食系统的一个组成部分，处于这一发展的最前沿。自2016年以来，研究已经通过使用动物的鬼脸量表来测量负面影响来确定面部动作单位的变化，但是在牲畜中使用鬼脸量表存在局限性。例如，个体的表达和疼痛的程度在动物之间可能会有所不同，这是很难区分的。由于这些原因和其他原因，在一些牲畜物种（如猪）中用于疼痛评估的鬼脸量表尚未得到充分验证。我的研究项目是独一无二的，因为它是唯一一个研究牲畜微行为生物识别技术的项目，作为用于精准农业的自动化技术平台的一部分。我的研究计划的总体目标是更好地了解牲畜微行为生物识别技术的定性和定量差异如何被纳入自动化系统，并与机器学习相结合，用于各种研究和农场应用。除了使用自动红外热成像平台来检测微行为生物特征以及作为早期疾病检测的一部分的热特征外，我的研究还使用了3D运动学运动捕捉技术来更好地了解微行为的生物力学。例如，在发情期表现出的微行为的运动学，证实了微运动更好地检测生物功能的变化。然而，一个尚未探索的领域是使用微行为生物识别技术来评估牲畜的情感状态和面部表情。拟议研究的重点是回答动物福利中的一个新问题：微面部表情和其他微行为生物识别技术是否是牲畜疼痛或其他情感状态的副现象？如果是这样的话，潜在的生理机制是什么？本研究的短期目标是：1）表征与积极和消极情感状态相关的微面部表情，2）使用三维运动学分析，热成像和基因组分析来确定微面部表情作为情感生物识别的生理机制（途径）。这项拟议的研究将阐明生理学在驱动微面部表情以响应猪所经历的不同类型的刺激方面所起的作用，不仅致力于通过使用疼痛鬼脸量表来解决当前问题，而且还表明随着加拿大农业部门越来越多地采用精密系统，可以使用自动化集成机器学习来完成。