Our Idea for Creating Reliefs

We use a novel combinition of Machine Learning and Image Processing to create 3D representations of the artist's world. This allows us to create 3D printed tactile boards that folks with visual impairments can use to experience the world of art.

The Great Wave off Kanagawa

We will use Katsushika Hokusai's masterpiece The Great Wave off Kanagawa to illustrate our approach.

Depth Map

We start with getting a depth map of the painting via at DTP model - which gets us the depth of each pixel in the painting.

Canny Edge Detection

While very useful, the depth map does not have the finer details of the painting like textures or facial features. We use the popular Canny Edge detection algorithm to get these details

Combined Model

We then do a weighted combination the depth and feature maps to get a rich representation of the world in the painting.

Make it 3D Printable

This model is then converted in into an STL file that can be printed using any 3D printer.