Deepic

• 

  GANs

Inspiration

As William Shakespeare once said, "The eyes sees all, but the mind shows what we want to see." Our team seek to build an application that allow people to see the world differently by exchanging the styles of two views.

What it does

Users upload a style image and a content image(.jpg or .png). Our deep learning engine will intelligently transfer the style from the style image to the content image!

How I built it

Inspired by the recent advances in Deep Reinforcement, Generative Adversarial Network (GAN) improves itself through self-competition, featuring a live battle between two deep neural networks. Our Generator generates an image conditioned on the style of the given input.

Challenges I ran into

Network Tuning Lack of Computational Resource Lack of time

Accomplishments that I'm proud of

Fully functional project.

What's next for Deepic

Make it more powerful, maybe incorporate human-face detection into it for better style-changes for pictures with faces

Built With

• deep-learning
• gans
• html5
• pytorch