2025 Volume 61 Issue 10 Pages 471-478
1) In learning clinical skills, there is a problem that the correctness of the training is unknown by self-study. Therefore, we propose a method to recognize the behavior of learners during examination training to develop an examination training system that enables auto-scoring. 2) In this system, the transition of patient's posture during the examination was modeled by the state transition model with reference to the learning evaluation items. We developed a database of transition instructions based on the experiment. 3) The transition instructions include various representations used during the examination, such as those consisting of voice only or voice and motions. The system recognizes noun and verb in the learner's voice and direction vector in motions during the medical examination. 4) We evaluated the voice recognition rate, the result was 99%. And we evaluated the behavior recognition method consist of voice recognition and motion recognition, the result was 90%.
