ACADSTAFF UGM
| Title | : | Alignment Based Siamese Network Model for Face Verification |
| Author | : |
Rizky Agung Dwi Putranto (1) Wahyono, Ph.D. (2) |
| Date | : | 31 2019 |
| Keyword | : | Face recognition,verification,face alignment,convolutional neural network,siamese network Face recognition,verification,face alignment,convolutional neural network,siamese network |
| Abstract | : | Face verification, which is the subset of face recognition task, is one of the popular topics in machine vision. The task of face verification is to accept or reject the two image who claims to be same person. Face recognition, in general, has several problems. One of them is pose variations. With the alignment process, the pose variation problem is expected to be resolved. Convolutional Neural Network (CNN), on the other hand, is good for recognizing pattern and as a feature extractor in images. So, we build a siamese network based on CNN for verification task. The result of this study is the alignment process can increase the accuracy of the system. Accuracy of the system is 51.7% with no alignment and 68.9% when using alignment. |
| Group of Knowledge | : | Ilmu Komputer |
| Original Language | : | English |
| Level | : | Internasional |
| Status | : |
Published
|
| No | Title | Action |
|---|