Application of Deep Convolutional Generative Adversarial Networks to Generate Pose Invariant Facial Image Synthesis Data
Dublin Core
Title
Application of Deep Convolutional Generative Adversarial Networks to Generate Pose Invariant Facial Image Synthesis Data
Subject
face recognition; pose invariant; generative adversarial networks; deep convolutional generative adversarial
networks; hyperparameter tuning
networks; hyperparameter tuning
Description
Even though Artificial Intelligence is advancing, artificial intelligence can still find it difficult to solve problems that are easy
for humans to do but difficult for computers to describe, such as facial recognition. There are problems related to the existing
facial recognition model, namely the facial recognition model. The model is still unable to recognize facial shapes that are not
in a perfect state due to several factors. Among several factors, the most influencing factor is the position of the face. Therefore,
in this study, Deep Convolutional Generative Adversarial Networks (DCGAN) will be applied to generate fake image data with
varying face positions. This research will be carried out starting from collecting data, processing data, designing, and training
models, hyperparameter tuning, and lastly analyzing test results. Based on the results of hyperparameter tuning that carried
out sequentially, the best hyperparameter combination produced is 200 epoch, 0.002 Generator learning rate, 0.5 Generator
momentum/beta1, Adam as Generator optimizer, 0.0002 Discriminator learning rate, 0.5 Discriminator momentum/beta1, and
Adam as Discriminator optimizer. The hyperparameter combination gives a result with FID score of 74.05. Based on testing
with human observer, generated fake images has relatively good results, but there are still few bad fake image results
for humans to do but difficult for computers to describe, such as facial recognition. There are problems related to the existing
facial recognition model, namely the facial recognition model. The model is still unable to recognize facial shapes that are not
in a perfect state due to several factors. Among several factors, the most influencing factor is the position of the face. Therefore,
in this study, Deep Convolutional Generative Adversarial Networks (DCGAN) will be applied to generate fake image data with
varying face positions. This research will be carried out starting from collecting data, processing data, designing, and training
models, hyperparameter tuning, and lastly analyzing test results. Based on the results of hyperparameter tuning that carried
out sequentially, the best hyperparameter combination produced is 200 epoch, 0.002 Generator learning rate, 0.5 Generator
momentum/beta1, Adam as Generator optimizer, 0.0002 Discriminator learning rate, 0.5 Discriminator momentum/beta1, and
Adam as Discriminator optimizer. The hyperparameter combination gives a result with FID score of 74.05. Based on testing
with human observer, generated fake images has relatively good results, but there are still few bad fake image results
Creator
Jagad Nabil Tuah Imanda, Fitra Abdurrachman Bachtiar, Achmad Ridok
Source
http://jurnal.iaii.or.id
Publisher
Professional Organization Ikatan Ahli Informatika Indonesia (IAII)/Indonesian Informatics Experts Association
Date
October 2023
Contributor
Sri Wahyuni
Rights
ISSN Media Electronic: 2580-0760
Format
PDF
Language
English
Type
Text
Files
Collection
Citation
Jagad Nabil Tuah Imanda, Fitra Abdurrachman Bachtiar, Achmad Ridok, “Application of Deep Convolutional Generative Adversarial Networks to Generate Pose Invariant Facial Image Synthesis Data,” Repository Horizon University Indonesia, accessed February 3, 2026, https://repository.horizon.ac.id/items/show/10104.