Artificial Intelligence Technology in 3D Facial Reconstruction: An Approach to Reutilize 2D Standardized Images in Plastic Surgery.

[INTRODUCTION] Two-dimensional (2D) facial photographs serve as crucial data in plastic surgery, and their reuse can yield additional clinical information for reference purposes. The present study presents an approach for three-dimensional (3D) facial reconstruction, wherein the patient's standard facial photograph is imported into a reconstruction algorithm to generate a 3D facial model. This reconstructed model allows for comprehensive observation from various angles, providing accurate representation of the proportion and structural characteristics of the facial contour.

[METHODS] A total of 31 patients were randomly selected, and the frontal and lateral photographs of each patient's face were input into the algorithm for facial 3D reconstruction. The facial subunits were analyzed in terms of their proportions, angles, and lengths. The effectiveness of the reconstruction algorithm was evaluated by comparing the dimensions of each index in both frontal and lateral 2D standardized facial photographs, as well as 3D scans of the patient, along with frontal and lateral perspectives of 3D model generated by the reconstruction algorithm.

[RESULTS] The algorithm-reconstructed 3D model group showed significant correlations with both the standard photograph group and the 3D scan-reconstructed model group in each indicator, thereby indicating the precise reconstruction of patients' facial information through this reconstruction algorithm.

[CONCLUSION] The facial 3D reconstruction algorithm possesses the capability to transform 2D photographs into precise 3D digital models, exhibiting efficiency, convenience, speed, and accessibility while accurately capturing the distinctive attributes of facial subunit. This approach facilitates the reutilization of 2D facial photographs, thereby providing a genuine and invaluable tool for clinical practice.

[LEVEL OF EVIDENCE III] This journal requires that authors assign a level of evidence to each article. For a full description of these Evidence-Based Medicine ratings, please refer to the Table of Contents or the online Instructions to Authors   www.springer.com/00266 .