Scoping Review on Machine Learning Models for Detecting Salmonella Infection in Chicken Farms

Salmonella infections pose a significant risk to chicken farms. Literature on the application of Machine Learning (ML) for the detection of Salmonella infections is diverse and scattered, and mapping studies on the topic are limited. This scoping review explores the application of ML models for detecting Salmonella infection in poultry farming. A systematic search across PubMed, Scopus, IEEE Xplore, and Web of Science identified 37 relevant studies published between 2021 and 2025. Reported impacts of Salmonella outbreaks include economic losses, public health risks, animal welfare concerns, hospitalizations, and trade disruptions, with economic impacts accounting for the largest share (35%). The findings again reveal that most proposed solutions employ supervised learning techniques, with Convolutional Neural Networks (CNNs) being the predominant approach, particularly for image-based data such as faecal and hyperspectral imaging. The review highlights high success rates among deep learning models, with several achieving accuracies above 95%, though limitations remain in their reliance on large datasets and image-based data types. Notably, there is limited exploration of alternative ML approaches capable of integrating multimodal data such as sensor readings, behavioral indicators, or environmental parameters. This gap presents opportunities for future research to develop more versatile, explainable, and scalable ML models for early detection of Salmonella in chicken farming.