Date Thesis Awarded

5-2018

Access Type

Honors Thesis -- Access Restricted On-Campus Only

Degree Name

Bachelors of Science (BS)

Department

Computer Science

Advisor

Timothy A. Davis

Committee Members

James W. Deverick

Chun-yu Lu

Abstract

This thesis describes an algorithm that generates various ink bamboo paintings. First, a completely procedural model is used to generate the geometric shape of bamboos. The model uses a grammar-like approach that recursively generates new parts of the bamboo in a randomized manner. The random parameters are bounded by rules that simulate the natural form of bamboo.

The structure of the bamboo is represented line segments with directions. Various ink stroke sprites of stalk, branch, or leaf shapes are mapped to line segments, using reverse mapping and bilinear sampling to eliminate aliasing effects. The sprites are mapped in different degrees of transparency to simulate the effect of various shades of ink produced by changes in forces when using an ink brush. Finally, a seal is applied to sign the work and enhance the visual effect.

The algorithm is implemented in Python 3 and can be run on any computer with the imageio library installed. The output of the program is saved in a PNG image file, which can be used for various types of illustrations. This model is able to produce unique images during every run, and would significantly reduce human labor in painting stylistically similar artworks of ink bamboo paintings.

Download

Share Feedback

On-Campus Access Only

Share

COinS