Visual Effect Generation Via Deep Convolutional Generative Adversarial Network

Visual effect is important as it can make a film/video look more realistic. However, generating visual effect is costly and time consuming. In this technology era, the use of neural networks becomes more and more common. A lot of expertise starts to use neural networks to generate images, videos, and others. There are many types of artificial neural networks having different purposes and functions. From these neural networks, it is found that generative adversarial neural network (GAN) has the best performance to generate data and images. In addition, applying convolution ideas makes GAN more reliable and stable. Thus, in this research, a new model of deep convolutional generative adversarial network (DCGAN) to generate visual effect is proposed.