BirthARssistant: a childbirth delivery training simulator integrating microsoft HoloLens2, 3D-printing and electromagnetic tracking.

PURPOSE: Manual assistance during vaginal delivery requires precise hand coordination to guide fetal emergence while protecting the maternal perineum. However, these hands-on skills cannot be acquired through theoretical instruction alone, while conventional mannequin-based training requires continuous expert supervision and relies largely on subjective feedback, limiting opportunities for standardized practice. This work presents BirthARssistant, an augmented reality simulator that integrates obstetric mannequins, Microsoft HoloLens 2, 3D-printing, electromagnetic tracking and real-time hand tracking to provide measurable guidance during childbirth training. METHODS: The simulator registers and overlays virtual anatomical models onto obstetric mannequins using QR markers and real-time electromagnetic tracking. Tracking data is processed in 3D Slicer and streamed to Unity through OpenIGTLink. Reference hand postures were recorded from experts for seven stages of vaginal delivery and compared with participants' performance. Eleven participants completed pre- and post- training assessments based on hand placement accuracy, positional and angular errors, questionnaires, and paired statistical analyses. RESULTS: After training, correct hand positioning increased by 57.1% and correct dominant-hand use increased by 30.6%, with statistically significant improvements in both variables. Mean positional error decreased from 64.7 ± 30.1 to 47.1 ± 36.2 mm, while mean angular error decreased from 11.9 ± 11.1° to 8.4 ± 8.2°. Questionnaire responses indicated favorable perceptions of the simulator's usability and educational value. CONCLUSIONS: BirthARssistant provides an interactive and measurable approach to obstetric simulation by combining mannequin-based training with AR guidance and real-time motion tracking. The overall findings suggest that the simulator can help reduce inter- and intra-participant variability in performance, support error identification and facilitate skill acquisition by reinforcing the correct sequence and execution of the required maneuvers.