This Repository contains things I have created in order to study Generative Adversarial Networks(GAN)[1].
이 레포지토리에는 제가 GAN을 공부하기 위해서 만들었던 것들을 모아두었습니다.

Currently, there are two projects related to GAN, and both projects used Facial Expression Recognition (FER) datasets as inputs.

In this project, I wanted to find out the relationship between the input dataset and the performance of GAN. With Two FER datasets, I tried to analyse the properties of these datasets using different techniques. One of these techniques is to creaste histogram graphs of the dataset to see number of Mixture of Gaussians (MoG) that the dataset contain. Details can be seen in the project's repository.

In this project, I tried to modify the CatGAN [2] with various state-of-the-art generative, and classification networks. Details can be seen in the project's repository.

[1] I. J. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, Y. Bengio. Generative Adversarial Nets. Adv Neural Inf Process Syst, 2014, 27.
Github link

[2] J. T. Springenberg, “Unsupervised and Semi-Supervised Learning with Categorical Generative Adversarial Networks,” ICLR 2016.