Automated Cephalometric Landmarking System


  • To design and develop an automated approach for identifying cephalometric landmarks for the ease of doctors to assist in surgery.


  • The landmarking technique developed uses Convolutional Neural Network and Image Processing to achieve great accuracy. 
  • The model takes X-Ray radiographs as input. The output generated is the same radiograph with 19 landmarks annotated on it. 
  • Uses data augmentation to drastically increase the size of training data available. 
  • Does not require any manual intervention to generate results.


It can be used by doctors to annotate the cephalometric landmarks which will assist them in surgery planning.

© Design Innovation All Rights Reserved