Hybrid Deep Learning with Multi-Level Context for Pain Assessment Using Physiological Signals

A holistic way using a deep learning-based approach for automated pain recognition using physiological signals. Unlike traditional methods reliant on expert knowledge for feature extraction, this approach employs multi-level contextual information for superior classification. Experiments are conducted using the Bovid-Heat-Pain and Mopani datasets illustrates multi-level contextual learning improves classification accuracy significantly, achieving an average accuracy of 87.8%. These findings establish the effectiveness of deep-learning in automating pain recognition systems.