Implementation of an Attention Mechanism Model for Facial Beauty Assessment Using Transfer Learning
An important consideration in medical plastic surgery is the evaluation of the patient’s facial symmetry. However, because facial attractiveness is a slightly individualized cognitive experience, it is difficult to determine face attractiveness manually. This study aimed to train a model for assessi...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-03-01
|
| Series: | Diagnostics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2075-4418/13/7/1291 |
| _version_ | 1797608161695236096 |
|---|---|
| author | Chao-Tung Yang Yu-Chieh Wang Lun-Jou Lo Wen-Chung Chiang Shih-Ku Kuang Hsiu-Hsia Lin |
| author_facet | Chao-Tung Yang Yu-Chieh Wang Lun-Jou Lo Wen-Chung Chiang Shih-Ku Kuang Hsiu-Hsia Lin |
| author_sort | Chao-Tung Yang |
| collection | DOAJ |
| description | An important consideration in medical plastic surgery is the evaluation of the patient’s facial symmetry. However, because facial attractiveness is a slightly individualized cognitive experience, it is difficult to determine face attractiveness manually. This study aimed to train a model for assessing facial attractiveness using transfer learning while also using the fine-grained image model to separate similar images by first learning features. In this case, the system can make assessments based on the input of facial photos. Thus, doctors can quickly and objectively treat patients’ scoring and save time for scoring. The transfer learning was combined with CNN, Xception, and attention mechanism models for training, using the SCUT-FBP5500 dataset for pre-training and freezing the weights as the transfer learning model. Then, we trained the Chang Gung Memorial Hospital Taiwan dataset to train the model based on transfer learning. The evaluation uses the mean absolute error percentage (MAPE) value. The root mean square error (RMSE) value is used as the basis for experimental adjustment and the quantitative standard for the model’s predictive. The best model can obtain 0.50 in RMSE and 18.5% average error in MAPE. A web page was developed to infer the deep learning model to visualize the predictive model. |
| first_indexed | 2024-03-11T05:39:35Z |
| format | Article |
| id | doaj.art-26f1ccbecb434421be37dcc8785f9c9c |
| institution | Directory Open Access Journal |
| issn | 2075-4418 |
| language | English |
| last_indexed | 2024-03-11T05:39:35Z |
| publishDate | 2023-03-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Diagnostics |
| spelling | doaj.art-26f1ccbecb434421be37dcc8785f9c9c2023-11-17T16:30:32ZengMDPI AGDiagnostics2075-44182023-03-01137129110.3390/diagnostics13071291Implementation of an Attention Mechanism Model for Facial Beauty Assessment Using Transfer LearningChao-Tung Yang0Yu-Chieh Wang1Lun-Jou Lo2Wen-Chung Chiang3Shih-Ku Kuang4Hsiu-Hsia Lin5Department of Computer Science, Tunghai University, Taichung City 407224, TaiwanDepartment of Computer Science, Tunghai University, Taichung City 407224, TaiwanDepartment of Plastic and Reconstructive Surgery, Craniofacial Research Center, Chang Gung Memorial Hospital, No. 123, Dinghu Rd., Guishan Township, Taoyuan City 333423, TaiwanDepartment of Tourism and Recreation Management, Hsiuping University of Science and Technology, No. 11, Gongye Rd., Dali District, Taichung City 412406, TaiwanImaging Laboratory, Craniofacial Research Center, Chang Gung Memorial Hospital, No. 123, Dinghu Rd., Guishan Township, Taoyuan City 333423, TaiwanImaging Laboratory, Craniofacial Research Center, Chang Gung Memorial Hospital, No. 123, Dinghu Rd., Guishan Township, Taoyuan City 333423, TaiwanAn important consideration in medical plastic surgery is the evaluation of the patient’s facial symmetry. However, because facial attractiveness is a slightly individualized cognitive experience, it is difficult to determine face attractiveness manually. This study aimed to train a model for assessing facial attractiveness using transfer learning while also using the fine-grained image model to separate similar images by first learning features. In this case, the system can make assessments based on the input of facial photos. Thus, doctors can quickly and objectively treat patients’ scoring and save time for scoring. The transfer learning was combined with CNN, Xception, and attention mechanism models for training, using the SCUT-FBP5500 dataset for pre-training and freezing the weights as the transfer learning model. Then, we trained the Chang Gung Memorial Hospital Taiwan dataset to train the model based on transfer learning. The evaluation uses the mean absolute error percentage (MAPE) value. The root mean square error (RMSE) value is used as the basis for experimental adjustment and the quantitative standard for the model’s predictive. The best model can obtain 0.50 in RMSE and 18.5% average error in MAPE. A web page was developed to infer the deep learning model to visualize the predictive model.https://www.mdpi.com/2075-4418/13/7/1291transfer learningdeep learningfacial attractiveness predictionattention mechanismvisualization |
| spellingShingle | Chao-Tung Yang Yu-Chieh Wang Lun-Jou Lo Wen-Chung Chiang Shih-Ku Kuang Hsiu-Hsia Lin Implementation of an Attention Mechanism Model for Facial Beauty Assessment Using Transfer Learning Diagnostics transfer learning deep learning facial attractiveness prediction attention mechanism visualization |
| title | Implementation of an Attention Mechanism Model for Facial Beauty Assessment Using Transfer Learning |
| title_full | Implementation of an Attention Mechanism Model for Facial Beauty Assessment Using Transfer Learning |
| title_fullStr | Implementation of an Attention Mechanism Model for Facial Beauty Assessment Using Transfer Learning |
| title_full_unstemmed | Implementation of an Attention Mechanism Model for Facial Beauty Assessment Using Transfer Learning |
| title_short | Implementation of an Attention Mechanism Model for Facial Beauty Assessment Using Transfer Learning |
| title_sort | implementation of an attention mechanism model for facial beauty assessment using transfer learning |
| topic | transfer learning deep learning facial attractiveness prediction attention mechanism visualization |
| url | https://www.mdpi.com/2075-4418/13/7/1291 |
| work_keys_str_mv | AT chaotungyang implementationofanattentionmechanismmodelforfacialbeautyassessmentusingtransferlearning AT yuchiehwang implementationofanattentionmechanismmodelforfacialbeautyassessmentusingtransferlearning AT lunjoulo implementationofanattentionmechanismmodelforfacialbeautyassessmentusingtransferlearning AT wenchungchiang implementationofanattentionmechanismmodelforfacialbeautyassessmentusingtransferlearning AT shihkukuang implementationofanattentionmechanismmodelforfacialbeautyassessmentusingtransferlearning AT hsiuhsialin implementationofanattentionmechanismmodelforfacialbeautyassessmentusingtransferlearning |