An Improved cGAN with Self-Supervised Guidance Encoder for Generation of High-Resolution Facial Expression Images
Tatsuya Hanano, Masataka Seo, Yen‐Wei Chen · 2023 IEEE International Conference on Consumer Electronics (ICCE) · 2023
The recent spread of smartphones and social networking services has increased the means of seeing images of human faces. Particularly, in the field of facial images, the facial expression conversion has already been realized using deep learning-based approaches. However, in the conventional deep learning-based methods, only low-resolution images can be generated due to limited computational resources. As a result, the generated images are blurry or aliasing. To solve low-resolution problem, we proposed a serial method to enhance the resolution of the generated facial images by combining a super-resolution network following the generative model in our previous work. In this paper, we propose a new model with two encoders that integrates a self-supervised guidance encoder into cGAN to further improve the accuracy of the generated results. We used structural similarity and peak signal-to-noise ratio as evaluation metrics and were able to improve image quality compared to our previous baseline model.