Optimizing texture representation in 3D medical models using an RGBD camera

Introduction: Generating and visualizing digital models that closely resemble reality can significantly enhance user experiences in various medical activities such as surgical planning or educational phases; Methods: This research presents a method to improve a 3D digital model, reconstructed from cone-beam computed tomography (CBCT) images, by incorporating its real coloured texture captured using an Intel D435 RGBD camera. This method is based on evaluating the oriented bounding boxes of both pointclouds and then applying the roto-translation matrix by imposing that the front faces of both boxes would coincide. Subsequently, the coloured digital models were computed; Results: The alignment error between the two 3D pointclouds, assessed using 30000 random points, indicates values of: 1.9±2mm on the x-axis, 1.2±1.7mm on the y-axis, and 1.6±2mm on the z-axis; Conclusions: This study presents a methodology for enriching a 3D model reconstructed from CBCT images with its real coloured textures obtained through an RGBD camera. This method could be used to create digital 3D models for medical, surgical and educational purposes.