Preoperative Computer Simulation in Rhinoplasty Using Previous Postoperative Images.

Providing realistic and useful preoperative counseling based on a surgeon's knowledge and experience. Using previous preoperative and postoperative patients' images to predict the postoperative result of a new query patient. After preprocessing for image standardization, facial landmarking was done using 68 points on the frontal view and 19 points on the profile view. Facial features were calculated and image retrieval was done based on similarity measurement between the query image's feature vector and database images' feature vectors. The nasal areas on the postoperative retrieved images were swapped to the corresponding region on the query patient's face at a unicenter tertiary hospital. We randomly selected the color profile photographs of 400 patients (360 women and 40 men) from the database of all rhinoplasty patients who had been successfully operated at Valiasr hospital from 2010 to 2018. The accuracy of this preoperative simulation was >80% in our pilot study in 20 patients. This system is fast, easy to handle, reliable, and accurate to simulate postoperative outcomes both in frontal and profile views. We believe that this system could not only improve more informative communication between surgeons and patients but also could facilitate the training of surgeons and residents.