Visual animal biometrics is an emerging research field of computer vision, pattern recognition, and cognitive science. Recently, cattle recognition has played a significant role in understanding, controlling and the outbreak of critical diseases, vaccination, production management, traceability, and ownership assignment of a livestock animal. The traditional animal recognition methodologies, such as ear-tagging, freeze-branding, ear-tattoos, embedded microchips, ear tips or notches-based, and electrical-based marking approaches, have been applied to recognize individual livestock animal. However, standard animal recognition procedures are invasive. The performance of conventional methods is not good due to their vulnerability to losses, easy duplication, and fraud of embedded tag number. These are major security issues and challenges for the identification of cattle throughout the world. Visual animal biometric systems are gaining more proliferations due to widespread applications to recognize individual cattle based on their primary biometric muzzle point image characteristics. This paper aims to provide a comprehensive review of cattle recognition and tracking from non-biometric recognition approaches (classical animal recognition methods) to visual animal biometric systems using muzzle point image pattern along with measurements and interpretations based on current state-of-the-art methods. Moreover, this paper demonstrates the basic deployment of the animal biometric system to uniquely identify the animals using their biometric characteristics. This study can hopefully encourage new multidisciplinary researchers and scientists to provide excellent efforts for the designing and development of adequate algorithms for solving the classification and recognition problems. The literature review is followed by the presentation of references for more details, incorporating applications and current trends.

视觉动物生物识别技术是计算机视觉，模式识别和认知科学的新兴研究领域。最近，牛的识别在了解，控制和爆发重大疾病，疫苗接种，生产管理，可追溯性以及家畜的所有权分配方面发挥了重要作用。传统的动物识别方法，如耳标，冷冻烙印，耳纹身，嵌入式微芯片，基于耳尖或缺口的标记以及基于电子的标记方法，已被用于识别个体牲畜。但是，标准的动物识别程序具有侵入性。常规方法的性能不佳，因为它们容易丢失，容易重复以及嵌入标签号的欺诈。这些是全世界鉴定牛的主要安全问题和挑战。视觉动物生物识别系统正获得广泛的应用，这是由于基于其主要生物识别口吻图像特征来识别单个牛的广泛应用。本文旨在提供对牛的识别和跟踪的全面综述，从非生物识别方法（经典动物识别方法）到使用枪口点图像模式的视觉动物生物识别系统，以及基于当前最新技术的测量和解释方法。此外，本文演示了动物生物识别系统的基本部署，以利用其生物识别特征来唯一识别动物。这项研究有望鼓励新的跨学科研究人员和科学家为设计和开发适当的算法以解决分类和识别问题做出出色的努力。文献综述之后，将提供更多参考资料，并结合应用和当前趋势。