The goal of this research was to generate cartoon art that is creative. The aim was to generate animated figures with unique postures that do not exist and are completely novel. For this, we proposed an unsupervised learning-based method by using Convolutional Neural Network (CNN) with GAN to generate cartoon art.
Usually, visual animators work with artists to create a storyline for an animated film. This process is tedious as, once the character is finalized, they have to create a new sketch of the characters for every frame in the film. This process is extremely time consuming and costly.
We proposed a methodology for generating creative cartoon art. The system generated 'Tom the cat' from the famous Tom and Jerry cartoon by looking at various existing images of Tom's character and learning about his posture/animation style.
This application can be used in Entertainment and Media sectors. It is useful for creating new postures and frames of characters, creating short animated films. It can be used to create different stories, advertisements, generating rhymes, music and paintings.
Check out the project video below.
Amit GawadeSenior Innovation Engineer
REDX WeSchool, Mumbai
Dr. Subodh DeolekarLead Research Engineer
Assistant Prof. Research & Business Analytics
at Welingkar Institute
Rohit PandharkarChief Lab Mentor
Vice President and Head of
Data Science at Mahindra Finance