COMPUTATIONAL-AIDED INVESTIGATION OF FACIAL STRUCTURES AUTOMATIC SEGMENTATION AND VOLUME ESTIMATION OF MAXILLARIS SINUSES: AUTOMATIC SEGMENTATION OF MAXILLARIS SINUSES

Introduction: Correct identification of anatomical structures in image examinations is a foundational aspect of medical practice, influencing diagnosis, treatment, patient safety, and overall healthcare quality. In the scope of facial anatomy, maxillary sinuses evaluation aid the diagnosis of several endonasal injuries. Aims: This work presents an automatic method to segment and create 3D models of maxillary sinuses from ConeBeam Computed Tomograph (CBCT). Materials and Methods: For all CBCT axial slices, regions possibly corresponding to maxillary sinuses were defined, according to the Hounsfield values of their points. A region growing process was performed, and the noise in the resultant areas was removed. Finally, the 3D volumes are generated. and measured (in cm3). The evaluation of the proposed method included 7 exams of six women and one man with ages from 30 to 60. Results: The obtained volumes were close to the average value denoted by previous works in medical literature. Also, a visual assessment comparing the CBCT data and the volume shape demonstrated the precision of the results achieved by the proposed method. Conclusion: The results (volume values and visual assessment) suggest the potential of this approach.