Inspiration
What is art created not through the lense of the human brain but through the body of something that is not alive? Exploring this topic I decided to create an image generator through AI and explore the depths of data and autonomous creativity
What it does
Through the use of neural networks, I was able to implement Generative Adversarial Networks and Variational Autoencoders to create the image. I made use of Open AI's training data. Basically the Generative Adversarial Networks gathers random noise to create the images and the discriminator differentiates between a variety of images that are real or ai generated. Over time from training the AI, it would be able to better create more realistic images than ones that are considered ai. The Varational Autoencoders were used to take in new input data and generate new data points. It takes care of an encoder network that describes the maps input to a latent space and through a decoder network back to images.
What we learned
I learned about the interesting field that was found between the lands of technology and creativity. Through neural networks such as VAEs(Varational Encoders) and GANs(Generative Adversarial Networks) I was able to discover the vast datasets that would eventually build the images that we describe
What's next for Art Generation
Possibly in the future I would like to learn about how accurate AI can get at creating images and how this can better help content creation, virtual reality, gaming, fashion design, medical imaging, product design, or architecture and real estate
Log in or sign up for Devpost to join the conversation.