51. Random Forest For Hijab Style Selection Based on Face Shape Using Morphological Facial Index
- Author
-
Aris Puji Widodo, K. N. Dinar Mutiara, and Wina Ratna Wati
- Subjects
Identification (information) ,Computer science ,business.industry ,Face (geometry) ,Feature extraction ,Pattern recognition ,Artificial intelligence ,Landmark point ,Style guide ,business ,Selection (genetic algorithm) ,Test data ,Random forest - Abstract
Face shape identification is needed to reference the hijab style selection. Many people don’t know their face shape. The mistake of choosing a hijab style makes the face look disproportionate. The morphological facial index model is used to identify the face shape by calculating the length and width of the face. Face Landmark point is used for facial feature extraction which is used as an attribute in the study, namely face height, face width, upper jaw width and lower jaw width. Various kinds of hijab styles and the absence of a hijab style guide make it difficult to choose a hijab style that suits your face shape. Machine learning is needed to create a hijab style selection system. Random Forest was chosen as the method for choosing the hijab style. The purpose of this study is the implementation of a random forest for the selection of hijab styles based on face shape using morphological facial index. This study used 129 images of female facial data which was split into 89% training data and 11% test data. The prediction system produces an accuracy value of 93% with 30 trees for pashmina, scarf and instant hijab styles. The use of the random forest method is proven to be able to predict the selection of hijab styles based on face shape using morphological facial index. Scikit-learn tool used to random forest modelling and evaluated the efficiency of our proposed method.
- Published
- 2021
- Full Text
- View/download PDF