Journal
COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE
Volume 200, Issue -, Pages -Publisher
ELSEVIER IRELAND LTD
DOI: 10.1016/j.cmpb.2021.105928
Keywords
Transfer learning; CNN; Facial symmetry; Deep learning; Data preprocessing
Categories
Funding
- Ministry of Science and Technology, Taiwan [MOST 108-2314-B-182A-001, MOST 108-2314-B-182-001, MOST 108-2314-B-182-002, MOST 108-2221-E-029-010, MOST 108-2745-8-029-007, MOST 108-2622-E-029-007-CC3]
- Chang Gung Memorial Hospital [CMRPG5F0181-2]
Ask authors/readers for more resources
This study introduced a convolutional neural network with a transfer learning approach for facial symmetry assessment based on 3-dimensional features to assist physicians in enhancing medical treatments. By transforming 3D features and applying data augmentation methods, a new model for evaluating facial symmetry was successfully trained, achieving satisfactory experimental results.
Orthognathic surgery (OGS) is frequently used to correct facial deformities associated with skeletal malocclusion and facial asymmetry. An accurate evaluation of facial symmetry is a critical for precise surgical planning and the execution of OGS. However, no facial symmetry scoring standard is available. Typically, orthodontists or physicians simply judge facial symmetry. Therefore, maintaining accuracy is difficult. We propose a convolutional neural network with a transfer learning approach for facial symmetry assessment based on 3-dimensional (3D) features to assist physicians in enhancing medical treatments. We trained a new model to score facial symmetry using transfer learning. Cone-beam computed tomography scans in 3D were transformed into contour maps that preserved 3D characteristics. We used various data preprocessing and amplification methods to determine the optimal results. The original data were enlarged by 100 times. We compared the quality of the four models in our experiment, and the neural network architecture was used in the analysis to import the pretraining model. We also increased the number of layers, and the classification layer was fully connected. We input random deformation data during training and dropout to prevent the model from overfitting. In our experimental results, the Xception model and the constant data amplification approach achieved an accuracy rate of 90%. (c) 2021 Elsevier B.V. All rights reserved.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available