Feature extraction: hand shape, hand position and hand trajectory path

Vision-based hand posture detection and tracking is an important issue for Human to Computer Interaction applications. The performance of recognition system fIrst depends on the process of getting effIcient features to represent pattern characteristics [1]. There is no algorithm which shows how to...

Full description

Saved in:
Bibliographic Details
Main Authors: Bilal, Sara Mohammed Osman Saleh, Akmeliawati, Rini
Format: Book Chapter
Language:English
Published: IIUM Press 2011
Subjects:
Online Access:http://irep.iium.edu.my/21640/1/Chapter_11.pdf
http://irep.iium.edu.my/21640/
http://rms.research.iium.edu.my/bookstore/default.aspx
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Universiti Islam Antarabangsa Malaysia
Language: English
Description
Summary:Vision-based hand posture detection and tracking is an important issue for Human to Computer Interaction applications. The performance of recognition system fIrst depends on the process of getting effIcient features to represent pattern characteristics [1]. There is no algorithm which shows how to select the representation or choose the features [2] so the selection of features will depend on the application. There are many different methods to represent 2-D images such as boundary, topological, shape grammar, description of similarity etc. [2-4]. Features should be chosen so that they are intensive to noise-like variation in pattern and keep the number of feature small for easy computation [5]. Hand posture shape features, motion trajectory feature and hand position with respect to other human upper body parts play an important role within the preparation stage of the gesture before recognition. In this chapter, features have been extracted from hand posture closed contours, hand posture trajectory and hand position has been identifIed. Algorithms have been developed for extracting these features after segmenting the head and the two hands. These extracted features can be attached to a recognizer such as Support Vector machine, Hidden Markov Model, etc. for hand gesture recognition.