Portal de Periódicos da CAPES

A Study on Static Hand Gesture Recognition Using Type-1 Fuzzy Membership Function

2017; IGI Global; Linguagem: Inglês

10.4018/978-1-5225-3129-6.ch005

2327-3461

Sriparna Saha, Amit Konar,

Robot Manipulation and Learning

The idea of this chapter is the use of Gaussian type-1 fuzzy membership functions based approach for automatic hand gesture recognition. The process has been carried out in five stages starting with the use of skin color segmentation for the isolation of the hand from the background. Then Sobel edge detection technique is employed to extract the contour of the hand. The next stage comprises of the calculation of eight spatial distances by locating the center point of the boundary and all distances are normalized with respect to the maximum distance value. Finally, matching based on Gaussian fuzzy membership function is used for the recognition of unknown hand gestures. This simple and effective procedure produces highest accuracy of 91.23% for Gaussian membership function and a time complexity of 2.01s using Matlab R2011b run on an Intel Pentium Dual Core Processor.