Animo

Inspiration

Current Problems in the anime industry: 1) Many frames are drawn manually ( Results in poor quality) 2) High frames per second output required for smooth animations 3) Pay of frame animators is low. Other tough parts include scene and background creation.

All these problems were my inspirations to solve this problem.

What it does

The ideal scenario: 1) Use Deep learning to convert real time acting to drawings. 2) Generate new backgrounds and scenes with only text descriptions. 3) Create manga sequences with video/text.

How we might built it

1) Pose estimation 2) Pix2Pix - Deep Learning 3) Generative Adversarial Networks (GANs)

Challenges we ran into

Might be compute and data majorly.

Accomplishments that we're proud of

What we learned

What's next for Animo

State of the Accuracy of deep learning models.

Built With

• python
• tensorflow