Posts Computational Image
Post
Cancel

Computational Image

This post was created in July 2021 and back-dated to match the approximate date of the content. See this post or group of posts for context and more info.

Backstory

In college, I took a Computational Image course that covered topics including image formation, feature extraction, object segmentation, recognition, and tracking, rendering, and multi-view geometry.

2011_feature_detection

For the final project for this class, I helped build out a multi-touch surface. The design for the multi-touch involved processing a real-time stream of infrared depth images coming from an Xbox Kinect and converting those image frames into features and their corresponding velocity. This was done by calculating a change in location of detected fingertip between sequential input frames. To demonstrate the functionality and test reliability of the design, I built a four-player game inspired by Osmos. Specifically, when a fingertip collided with an in-game object, the speed and direction that object was traveling would change accordingly. Below are videos demonstrating both the simulated gameplay and in-person gameplay. The program can be downloaded from here, circles.jar.

Illustrations

These are additional projects from this class that I resurrected based on converting text to colors and exploring color patterns and interactions.

test-to-colorcolor-compare
This post is licensed under CC BY 4.0 by the author.
Contents