AI Assisted Video Generation for Global Medical Physics Education: Aapm Virtual Training Resource Workgroup Experience
Abstract
Purpose
The Virtual Training Resource Work Group (VTRWG), operating under the International Council's Global Clinical Education and Training Committee, is charged with developing focused, practical clinical training programs. These programs are delivered virtually through interactive videos or in-person sessions, with the goal of establishing a sustainable platform for clinical education. The work group also partners with other AAPM committees for collaborative training opportunities.
Methods
To rank the need, we surveyed the global medical physics community and identified the most requested and pertinent training topics. Responses were scored and weighted by complexity, and three priority topics were selected. We drafted scripts for each and solicited reviews from the authoritative committees overseeing these topics. We then evaluated commercial services that leverage closed-source large language models to segment scripts into scene-specific text chunks, generate corresponding visual prompts, apply a consistent artistic style, and produce short video clips using text-to-video generative AI models.
Results
Between 2024 and 2025, we developed two training videos: Dosimetric Leaf Gap measurements in a water tank, Chair Test for DLG Verification. Each script and video underwent multiple rounds of review involving the VTRWG, GCETC, International Council, and relevant MPPG groups. The DLG measurement video has been posted to YouTube which garnered 2771 views and the Chair Test video is in final review.
Conclusion
AI-assisted video generation substantially reduces the time and specialized skill required to produce professional-quality training content. We plan to formalize a production pipeline along with policies and procedures that will enable any AAPM work group or task group to request hands-on training videos supporting their recommendations for the medical physics community.