
On October 10, 2026, at 3:00 PM
Speaker: Univ.-Prof. Dr. Dr. Ralf Smeets I Hamburg, DE
Objective:
The treatment of patients with facial defects, such as those caused by congenital malformations, presents a particular challenge not only for an interdisciplinary treatment team—comprising oral and maxillofacial surgeons and prosthodontists—but also for the patients themselves. When congenital or acquired defects in the head and neck region cannot be treated through surgical intervention—particularly following cancer or trauma—the existing defect must be reconstructed using prostheses. These prostheses not only conceal the defect but are also intended to be aesthetically sophisticated, thereby facilitating the patient’s psychosocial reintegration. Such an epithesis often represents the last chance for patients, for example, following chemotherapy. The working group’s goal was to capture and convert suitable data using AI to generate a detailed, lifelike 3D image for epithesis modeling.
Materials/Methods:
Various 3D scanning technologies were investigated to identify the most suitable method for rapid data acquisition without ionizing radiation. The target scan time was 30–60 seconds with a maximum radiation exposure of 200 μSv. A protocol to support data acquisition and ensure sufficient image quality was developed in collaboration with the project partners. This includes performing scans in a dedicated room with optimal lighting conditions, maintaining sufficient distance from the patient during the scanning process, and minimizing patient movement. Additionally, a method was developed to generate a 3D depth map of the scanned area, which enabled subsequent modeling.
Results:
Initial tests with selected technologies have shown that the target data acquisition time is achievable. Radiation exposure remains well below the limit. Point accuracy: The required point accuracy of at least 100 μm for capturing fine textural details, such as pores and/or wrinkles, has been successfully validated. Methods are currently being developed to accurately capture fine texture details on the surface of the face and neck, including the correct representation of cavities in the 3D model. The ideal peripheral lighting according to CIE Standard D65 was investigated. This revealed that under “standard clinical lighting” in the examination room, without special lighting measures, the defect scan can be performed with sufficient accuracy. A color depth of at least 24 bits per pixel was determined to be necessary to precisely map realistic skin discoloration. This parameter can be technically achieved within the framework of modern scanning methods. Using AI algorithms, we have developed and successfully implemented algorithms for noise filtering, gap filling, and normalization of the captured data. These steps are crucial for bringing the data into a uniform coordinate system and improving quality.
Conclusion:
This project focuses on automating the workflow for capturing and processing 3D data. So far, we have established a stable pipeline; however, there are still challenges in integrating user-friendly interfaces to make the system’s complexity more accessible to non-technical users. One of the biggest challenges lies in the quality of the captured data, which is often compromised by noise and artifacts. Additionally, processing large point clouds requires significant computational resources, which can lead to longer processing times. Developing a user-friendly interface that makes the complex functions accessible is also a challenge.
Conflict of Interest: We thank the Federal Ministry for Economic Affairs and Climate Action for its financial support (Grant No.: KK5180710IE3).


At the end of the congress, we will enjoy an enjoyable evening together on board of the “Weisse Flotte”.
Only a few tickets left

