Comprehensive Survey on Cattle Identification Approaches: From Traditional to Deep Learning Aspects

DOI: https://doie.org/10.1213/Jbse.2024126969

Meghna Luthra, Meghna Sharma, Poonam Chaudhary

Keywords:

Cattle identification, Animal husbandry, Machine learning, Artificial intelligence, Deep learning, Support Vector Machine, Computer vision

Abstract:

Cattle identification is essential for distinguishing individual animals, enabling effective tracking of disease progression, vaccination management, production monitoring, ensuring traceability, and establishing ownership. Recently, there has been a notable shift toward the automation of cattle identification, largely driven by advancements in machine learning (ML) and computer vision technologies. This paper provides a comprehensive review of various aspects of cattle identification techniques in the context of animal husbandry, exploring multiple facets of the field. It examines tag-based approaches that focus on hardware solutions, DNA feature-based methods emphasizing genetic identification, and visual feature-based approaches leveraging biometric data. Additionally, the paper discusses machine learning-based methods, deep learning techniques, and integrated learning approaches that combine both ML and Deep learning (DL) strategies for enhanced identification accuracy. Through this structured analysis, the paper aims to highlight the strengths and applications of each method in the advancement of cattle identification practices. Furthermore, the study investigates the standard image and video based datasets utilized in these models, emphasizing their characteristics, scale, and relevance to real-world applications. A thorough meta-analysis of key performance metrics is conducted to evaluate the effectiveness of different cattle identification approaches. Through this analysis, the paper aims to elucidate the strengths and limitations, offering valuable insights for future research and practical implementations in cattle identification.

PDF