How does thermal imaging compare with other pose estimation techniques? A survey on gait analysis in healthcare applications with a focus on knee osteoarthritis

Artificial Intelligence (AI) and Machine Learning (ML) have revolutionised numerous fields by enabling systems to analyse complex patterns and make accurate predictions from large datasets. In healthcare in particular, these technologies have been significantly employed to detect, diagnose, and monitor health conditions with minimal human intervention. ML models, especially those with deep learning techniques, have demonstrated strong potential in identifying relationships in data, making them ideal for applications like gait analysis and musculoskeletal disorder detection. This survey reviews AI and ML-based methods for identifying and classifying gait patterns mainly associated with musculoskeletal disorders such as knee osteoarthritis (OA), a chronic condition and one of the leading causes of disability in the UK. A central focus is on marker-based and markerless human pose estimation techniques, which enable precise posture and movement analysis with and without physical markers. Alongside this, advanced technologies such as thermal camera sensors are examined for their role in capturing gait-related features in a non-invasive and privacy-preserving manner. Finally, this survey provides a comprehensive overview of the field by bridging AI innovation with healthcare needs and its potential for improving the early detection and management of musculoskeletal disorders.