Recognizing expressive movement in arm gestures using Hidden Markov Models

Amanzadeh, Mohammad

Permalink

https://hdl.handle.net/2142/72866 Copy

Description

Title

Recognizing expressive movement in arm gestures using Hidden Markov Models

Author(s)

Amanzadeh, Mohammad

Issue Date

2015-01-21

Director of Research (if dissertation) or Advisor (if thesis)

Garnett, Guy E.

Department of Study

Industrial&Enterprise Sys Eng

Discipline

Systems & Entrepreneurial Engr

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

M.S.

Degree Level

Thesis

Keyword(s)

• Gesture Recognition
• Laban Movement Analysis
• Expressive Movement
• Hidden Markov Models

Abstract

Movement is one of the most basic human skills that used for communicating and interacting with the environment. Although we have an intuitive understanding of our gestures, it is hard to explain their quality. One would describe the human gestures as a collection of various actions for performing different tasks. True, but it does not explain how the tasks are performed, which is essential for having a more natural representation of movement. In this work we use Laban Movement Analysis (LMA), which is an analytic and experiential system for interpreting human body movement, to understand the expressive aspects of human gestures and we try to recognize them in hand movement using a supervised learning method with Hidden Markov Models (HMMs),; We first define the weight, space, and time characteristics of movement, which are described as the Basic Effort Factors (BEF) in LMA and we construct a classifier for each BEF using HMMs. We use a Microsoft Kinect to capture the body movement and try to recognize the quality of each BEF in hand gestures. Various preprocessing are done on the motion data to extract features that can describe the movement qualities. We use a windowing technique to segment the gestures into smaller sequences and allow for continuous gesture recognition. Various experiments are done to identify the optimal features and the parameters of each model, and we also address different problems in implementing the models.; Our research showed promising results in recognizing and distinguishing the BEFs of hand movement. The importance of this research is in developing systems that require deeper and more natural understanding of body movement, such as systems for recognizing musical gestures and dance, or producing computer animation.

Graduation Semester

2014-12

Permalink

http://hdl.handle.net/2142/72866

Copyright and License Information

Copyright 2014 Mohammad Amanzadeh

Owning Collections