Data for A Vision\u2013Depth Fusion Machine Learning Framework for Forward Head Posture Detection

Forward head posture (FHP) is a prevalent ergonomic concern during prolonged computer use, associated with musculoskeletal discomfort and long-term health implications. This study developed a compact, portable system that integrates time-of-flight depth sensing with camera-based imaging, in which the captured images are processed by a convolutional neural network to facilitate real-time monitoring of FHP. Posture data were collected from 100 participants using the device, and a lightweight multilayer perceptron (MLP)-based fusion algorithm was trained to classify FHP by integrating depth and vision features. Integrating depth and vision features at the feature level yielded a substantial improvement in classification accuracy over using either modality alone. The proposed model achieved an average accuracy of 0.9054, and an AUC of 0.9609. These results highlight the potential of this portable framework for real-time posture monitoring during computer use, offering a promising tool for ergonomic interventions. The proposed framework provides a cost-effective and scalable method for the early detection of musculoskeletal risk factors, thereby supporting preventive strategies in occupational and biomedical health.