Human-Centered Input Recognition Algorithms

Are you interested in natural user interaction? Do you want to learn more about how computer systems recognize and interpret user input in “natural” modalities, like touch, gesture, speech, and whole-body motion?​ Sign up now for a special topics tech elective being offered this Spring 2022 by Dr. Lisa Anthony, called “Human-Centered Input Recognition Algorithms.”

This course will cover typical approaches in recognition of input in these modalities that are informed by what we know about human input behaviors. This semester the emphasis will be on gesture-based recognition, including 2D surface gestures and 3D hand and finger gestures. Selected algorithms that will be covered may include: Wobbrock et al’s $1 recognizer, Vatavu et al’s $P recognizer, Taranta et al’s Jackknife 3D recognizer, Kratz & Rohs $3 3D recognizer, among others. Class structure will be in a project-based seminar format, in which we will discuss in-class weekly readings of the research papers that introduced these algorithms. Students will implement at least one of these algorithms and test it online in live demos and offline on sample data. Students will also extend at least one of these algorithms and test it in the same ways.

This course is scheduled for Thursdays periods 10-E1 (5:10-8:10pm) in NSC 0215 in Spring 2022.

If you are interested in this course, please see this draft syllabus from the last time it was taught in Spring 2019.

$P

$P 1-to-1 point matching between a template (light green)
and candidate (dark green) gesture [Vatavu et al, ICMI 2012].