|
|
Model Adaptation in Person Identification
Stratil, Jan ; Špaňhel, Jakub (referee) ; Hradiš, Michal (advisor)
This thesis deals with facial recognition using convolutional neural networks and with their current problems, which are pose, lighting and expression variance. It summarizes existing approaches, architectures and most recent loss functions. Further it deals with methods for rotating faces using GAN networks. In this thesis 3 neural networks are designed and trained for facial recognition. The best of them achieves 99.38% accuracy on LFW dataset and 88.08% accuracy on CPLFW dataset. Next face rotation network PCGAN is designed, which can be used for face frontalization or data augmentation purposes. This network is evaluated on Multi-PIE dataset and using the face frontalization it increases identification accuracy.
|
|
|
Face Image Frontalization Application
Tichý, Filip ; Malinka, Kamil (referee) ; Goldmann, Tomáš (advisor)
This work focuses on implementing an application for face frontalization using the CFR-GAN project and rotating the 3D face model followed by rendering. The aim of this work is to evaluate the impact of the application on face recognition accuracy based on the Fidentis dataset. The results are presented in the form of box plots, which depict the Euclidean distances between the generated frontalized images and the real images. It was found that when frontalizing using the rotation of a 3D model from high angles of rotation, the success of facial recognition process increases. Conversely, when frontalizing using the Complete Face Recovery GAN projekt, the recognition success signiĄcantly decreases. The VGG Face algorithm was used for comparing the images. The entire application is implemented in Python using commonly available libraries.
|
|
|
Model Adaptation in Person Identification
Stratil, Jan ; Špaňhel, Jakub (referee) ; Hradiš, Michal (advisor)
This thesis deals with facial recognition using convolutional neural networks and with their current problems, which are pose, lighting and expression variance. It summarizes existing approaches, architectures and most recent loss functions. Further it deals with methods for rotating faces using GAN networks. In this thesis 3 neural networks are designed and trained for facial recognition. The best of them achieves 99.38% accuracy on LFW dataset and 88.08% accuracy on CPLFW dataset. Next face rotation network PCGAN is designed, which can be used for face frontalization or data augmentation purposes. This network is evaluated on Multi-PIE dataset and using the face frontalization it increases identification accuracy.
|
guest ::
